<?xml version="1.0" encoding="utf-8"?>
<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Publishing DTD v1.1 20151215//EN" "JATS-journalpublishing1.dtd">
<article article-type="research-article" dtd-version="1.1" specific-use="sps-1.9" xml:lang="es" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">
<front>
<journal-meta>
<journal-id journal-id-type="publisher-id">teri</journal-id>
<journal-title-group>
<journal-title>Teor&#x00ED;a de la Educaci&#x00F3;n. Revista Interuniversitaria</journal-title>
<abbrev-journal-title abbrev-type="publisher">TERI</abbrev-journal-title>
</journal-title-group>
<issn pub-type="ppub">1130-3743</issn>
<issn pub-type="epub">2386-5660</issn>
<publisher>
<publisher-name>Ediciones Universidad de Salamanca</publisher-name>
</publisher>
</journal-meta>
<article-meta>
<article-id pub-id-type="publisher-id">teri.31821</article-id>
<article-id pub-id-type="doi">10.14201/teri.31821</article-id>
<article-categories>
<subj-group subj-group-type="heading">
<subject>Monogr&#x00E1;fico</subject>
</subj-group>
</article-categories>
<title-group>
    <article-title>Hacia un marco ético de la inteligencia artificial en la educación</article-title>
<trans-title-group>
<trans-title xml:lang="en"><italic>Towards an Ethical Framework of Artificial Intelligence in Education</italic></trans-title>
</trans-title-group>
</title-group>
<contrib-group>
<contrib contrib-type="author" corresp="yes">
<contrib-id contrib-id-type="orcid">https://orcid.org/0000-0002-2379-657X</contrib-id>
<name>
<surname>ALONSO-RODR&#x00CD;GUEZ</surname>
<given-names>Ana Mar&#x00ED;a</given-names>
</name>
<xref ref-type="aff" rid="aff1-ES"/>
<xref ref-type="corresp" rid="c1-ES"/>
</contrib>
<aff id="aff1-ES">
<institution content-type="original">Centro de Investigaci&#x00F3;n de Filosof&#x00ED;a de la Ciencia y la Tecnolog&#x00ED;a (CIFYT). Universidade de A Coru&#x00F1;a. (Campus de Ferrol). Espa&#x00F1;a.</institution>
<institution content-type="orgname">Centro de Investigaci&#x00F3;n de Filosof&#x00ED;a de la Ciencia y la Tecnolog&#x00ED;a (CIFYT)</institution>
<institution content-type="orgdiv1">Universidade de A Coru&#x00F1;a</institution>
<institution content-type="orgdiv2">Campus de Ferrol</institution>
<country country="ES">Espa&#x00F1;a</country>
</aff>
</contrib-group>
<author-notes>
<corresp id="c1-ES"><email>a.alonsor@udc.es</email></corresp>
</author-notes>
<pub-date publication-format="electronic" date-type="pub">
<day>04</day>
<month>06</month>
<year>2024</year>
</pub-date>
<pub-date publication-format="electronic" date-type="collection">
    <day>01</day>
    <month>07</month>
<year>2024</year>
</pub-date>
<volume>36</volume>
<issue>2</issue>
<issue-id pub-id-type="doi">10.14201/teri.2024362</issue-id>
<fpage>79</fpage>
<lpage>98</lpage>
<history>
<date date-type="received">
<day>29</day>
<month>12</month>
<year>2023</year>
</date>
<date date-type="accepted">
<day>22</day>
<month>02</month>
<year>2024</year>
</date>
</history>
<permissions>
<copyright-statement>&#x00A9; 2024 Ediciones Universidad de Salamanca</copyright-statement>
<copyright-year>2024</copyright-year>
<license license-type="open-access" xlink:href="https://creativecommons.org/licenses/by-nc-nd/4.0/" xml:lang="es">
<license-p>Esta obra est&#x00E1; bajo una licencia internacional Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International</license-p>
</license>
</permissions>
<abstract>
<title>RESUMEN</title>
<p>Este art&#x00ED;culo reflexiona sobre el uso de la Inteligencia Artificial en educaci&#x00F3;n desde una perspectiva &#x00E9;tica. Lo hace desde un punto de vista externo, considerando su incidencia en los contextos educativos como vivero de los desaf&#x00ED;os &#x00E9;ticos y pol&#x00ED;ticos que encara la sociedad. Esto permite dimensionar su alcance y profundidad y proponer medidas para afrontarlos.</p>
<p>Sus <italic>objetivos</italic> se enfocan hacia las bases &#x00E9;ticas de la Inteligencia Artificial relacionada con la actividad educativa, buscando identificar: las oportunidades, riesgos asociados y su impacto &#x00E9;tico en educaci&#x00F3;n; y b) los principios &#x00E9;ticos que puedan guiar el desarrollo, despliegue y uso de estos sistemas inteligentes.</p>
<p>Para ello se realiz&#x00F3; un estudio cualitativo, apoyado en una <italic>metodolog&#x00ED;a</italic> de revisi&#x00F3;n bibliogr&#x00E1;fica de: (i) trabajos acad&#x00E9;micos sobre sus usos actuales y riesgos potenciales de la Inteligencia Artificial; y (ii) un an&#x00E1;lisis comparativo de distintos c&#x00F3;digos &#x00E9;ticos, explorando la convergencia de principios aplicables a Sistemas de Inteligencia Artificial en los contextos educativos.</p>
<p>Los resultados obtenidos, en primer lugar, sit&#x00FA;an los problemas identificados en la tradici&#x00F3;n &#x00E9;tica, cuestionando la proliferaci&#x00F3;n de subdominios de la disciplina. Se indaga, despu&#x00E9;s, la posibilidad de un marco &#x00E9;tico unificado que evite la superposici&#x00F3;n de principios para cada dominio espec&#x00ED;fico. Se constata la utilidad de un marco ampliamente reconocido e influyente, cuyos principios se adaptan bien a los desaf&#x00ED;os de la educaci&#x00F3;n.</p>
<p>Se concluye se&#x00F1;alando las l&#x00ED;neas en las que se debe avanzar en la investigaci&#x00F3;n: (I) fundamentaci&#x00F3;n &#x00E9;tica y regulaci&#x00F3;n normativa para el desarrollo y uso de la Inteligencia Artificial en educaci&#x00F3;n conforme a los principios seleccionados; y (II) definici&#x00F3;n del nuevo perfil profesional docente y sus implicaciones para la formaci&#x00F3;n inicial del profesorado.</p>
</abstract>
<trans-abstract xml:lang="en">
<title>ABSTRACT</title>
<p>This article reflects on the use of Artificial Intelligence in education from an ethical perspective. It does so from the external point of view, considering its impact on educational contexts, as a breeding ground for the ethical and political challenges that society faces. This allows us to measure its scope and depth and propose actions to address them.</p>
<p>Its <italic>objectives</italic> focus on the ethical bases of Artificial Intelligence related to educational activity, seeking to identify: a) the opportunities, associated risks and its ethical impact on education; and b) the ethical principles that should guide the development, deployment and use of these intelligent systems.</p>
<p>For this purpose, a qualitative study was carried out, supported by a bibliographic review <italic>methodology</italic>: (i) academic works on current uses and potential risks of Artificial Intelligence and (ii) a comparative analysis of different ethical codes, exploring the convergence of principles applicable to Artificial Intelligence Systems in educational contexts.</p>
<p>The results obtained first place the identified problems in the ethical tradition and question the proliferation of subdomains of the discipline. The possibility of a unified ethical framework that avoids the overlap of principles for each specific domain is then investigated. The usefulness of a widely recognized and influential framework is confirmed, whose principles adapt well to the ethical challenges of education.</p>
<p>It concludes by indicating the lines in which research should advance (I) ethical foundation and normative regulation for the development and use of Artificial Intelligence in education in accordance with the selected principles; and (II) definition of the new teaching professional profile and its implications for initial teacher training.</p>
</trans-abstract>
<kwd-group xml:lang="es">
<title><italic>Palabras clave:</italic></title>
<kwd>educaci&#x00F3;n</kwd>
<kwd>inteligencia artificial</kwd>
<kwd>&#x00E9;tica</kwd>
<kwd>c&#x00F3;digos &#x00E9;ticos</kwd>
<kwd>marco &#x00E9;tico</kwd>
</kwd-group>
<kwd-group xml:lang="en">
<title><italic>Keywords:</italic></title>
<kwd>education</kwd>
<kwd>artificial intelligence</kwd>
<kwd>ethics</kwd>
<kwd>ethical codes</kwd>
<kwd>ethical framework</kwd>
</kwd-group>
</article-meta>
</front>
<body>
<sec id="sec-1-31821-ES">
<label><bold>1.</bold></label>
<title><bold>C<sc>ontexto actual</sc></bold></title>
<p>La digitalizaci&#x00F3;n creciente de las tareas habituales en los distintos sectores de la actividad humana ha generado cambios de car&#x00E1;cter informacional que modulan nuestra interacci&#x00F3;n con las Tecnolog&#x00ED;as de la Informaci&#x00F3;n y la Comunicaci&#x00F3;n (TICs), pero tambi&#x00E9;n c&#x00F3;mo se articulan las relaciones y los procesos sociales. Esto impulsa una &#x201C;reontologizaci&#x00F3;n&#x201D; del mundo a partir de la infosfera y nos sit&#x00FA;a en una nueva era: la hiperhistoria (<xref ref-type="bibr" rid="ref-18-31821-ES">Floridi, 2014</xref>), donde la informaci&#x00F3;n es el recurso fundamental, lo que nos convierte en vitalmente dependientes de las TICs.</p>
<p>La educaci&#x00F3;n es el medio que hace posible esta innovaci&#x00F3;n disruptiva y por eso es ahora &#x201C;a lo largo de la vida&#x201D;. Por ser la instancia que facilita la transici&#x00F3;n al recurso informativo, la Agenda 2030 tiene como uno de los objetivos del Desarrollo Sostenible (ODS 4) &#x201C;garantizar una educaci&#x00F3;n inclusiva, equitativa y de calidad y promover las oportunidades de aprendizaje durante toda la vida para todos&#x201D; (<xref ref-type="bibr" rid="ref-47-31821-ES">United Nations, 2019</xref>). La Inteligencia Artificial (IA) puede contribuir de manera eficaz a esos objetivos, pues ya ofrece numerosas oportunidades en los diversos &#x00E1;mbitos de la actividad social (la industria, la comunicaci&#x00F3;n, la salud, etc.).</p>
<p>Su incorporaci&#x00F3;n a la educaci&#x00F3;n impulsada desde el <italic>e-learning</italic> ha sido discreta, pero su uso est&#x00E1; en constante crecimiento incluso en los entornos educativos m&#x00E1;s convencionales. Su potencial para mejorar la educaci&#x00F3;n es inmenso, pero es preciso aprender a gestionar las numerosas implicaciones sociales, &#x00E9;ticas y deontol&#x00F3;gicas. Sobre todo, porque la normalizaci&#x00F3;n de la IA en los distintos &#x00E1;mbitos de la vida social conlleva un cambio inevitable al que la educaci&#x00F3;n est&#x00E1; obligada a responder.</p>
<p>Este &#x201C;discurso del cambio imperativo&#x201D; (<xref ref-type="bibr" rid="ref-6-31821-ES">Bearman <italic>et al.</italic>, 2023</xref>) supone poco realista prescindir de estas herramientas porque entra&#x00F1;an riesgos. Adem&#x00E1;s, ser&#x00ED;a contrario a los fines de la educaci&#x00F3;n, que incluyen preparar a los estudiantes para incorporarse al mundo en el que tendr&#x00E1;n que vivir, que es ya un entorno de digitalizaci&#x00F3;n e IA. Hay que pensar entonces en una <italic>Inteligencia Artificial para la Educaci&#x00F3;n</italic> en el contexto de una <italic>Educaci&#x00F3;n para la Inteligencia Artificial</italic>.</p>
<p>El impacto social de la IA en general plantea preocupaciones &#x00E9;ticas leg&#x00ED;timas respecto de su aplicaci&#x00F3;n en el dominio educativo. Por tratarse de un campo de especial riesgo &#x2014;y un inmejorable banco de pruebas para la resoluci&#x00F3;n de esta clase de problemas adecuadamente enfocados&#x2014; es necesaria una reflexi&#x00F3;n &#x00E9;tica. Puede plantearse desde una perspectiva &#x201C;interna&#x201D;, como hacen los cient&#x00ED;ficos que crean la IA, los tecn&#x00F3;logos que la implementan o los educadores que la utilizan. Pero aqu&#x00ED; se busca profundizar en el enfoque &#x201C;externo&#x201D;, que permite considerar la triple dimensi&#x00F3;n cient&#x00ED;fica, tecnol&#x00F3;gica y social, de los problemas y aportar criterios para valorar las posibles soluciones, los resultados que se esperan y las posibles consecuencias.</p>
<p>As&#x00ED;, los <italic>objetivos</italic> de este trabajo se orientan hacia las bases &#x00E9;ticas de la Inteligencia Artificial relacionada con la actividad educativa, buscando: a) identificar las oportunidades, riesgos e impacto &#x00E9;tico asociado a su incorporaci&#x00F3;n. Esto ayuda a comprender la necesidad de contar con principios &#x00E9;ticos; y b) concretar esos principios para disponer un marco &#x00E9;tico que regule el desarrollo, despliegue y uso de los sistemas inteligentes. Debe considerarse tambi&#x00E9;n su utilidad en la toma de decisiones pedag&#x00F3;gicas y para determinar las capacidades que necesitan docentes y estudiantes, cuando interaccionan con la IA, para prevenir sus riesgos.</p>
<p>Con el fin de articular una propuesta, se realiz&#x00F3; un estudio cualitativo apoyado en una metodolog&#x00ED;a de revisi&#x00F3;n bibliogr&#x00E1;fica de dos tipos de textos: (i) trabajos acad&#x00E9;micos sobre usos, riesgos y directrices &#x00E9;ticas en Inteligencia Artificial &#x2014; seleccionados en base a la tem&#x00E1;tica espec&#x00ED;fica, la autoridad de los autores y el impacto de los trabajos&#x2014; que confirmaron la necesidad de vincular las aplicaciones de IA a c&#x00F3;digos &#x00E9;ticos; y (ii) c&#x00F3;digos &#x00E9;ticos para sistemas inteligentes buscando consensos en torno a determinados principios. El n&#x00FA;mero y diversidad de los c&#x00F3;digos publicados recomend&#x00F3; un muestreo intencionalmente orientado a comparar la perspectiva de distintas partes interesadas: organizaciones internacionales, pa&#x00ED;ses implicados en la carrera por la IA y las industrias representativas del sector.</p>
<p>Dentro de este contexto, la investigaci&#x00F3;n sigue varios pasos que definen la estructura del art&#x00ED;culo. Primero, se sit&#x00FA;a el problema en el marco filos&#x00F3;fico-metodol&#x00F3;gico de las Ciencias de Dise&#x00F1;o. Segundo, se describen usos, oportunidades y riesgos de las herramientas de IA en educaci&#x00F3;n. Tercero, se constatan importantes implicaciones &#x00E9;ticas. A este respecto, una vez reconocida su continuidad con problemas de la tradici&#x00F3;n, se cuestiona la proliferaci&#x00F3;n de subdominios en la &#x00E9;tica. Cuarto, como resultado del an&#x00E1;lisis de distintos c&#x00F3;digos &#x00E9;ticos, se confirma la convergencia de principios que permiten proponer un marco unificado para la IA en educaci&#x00F3;n. Quinto, se concluye que este planteamiento requiere avanzar en dos direcciones de investigaci&#x00F3;n: (i) fundamentaci&#x00F3;n &#x00E9;tica y regulaci&#x00F3;n normativa para el desarrollo y uso de los Sistemas de Inteligencia Artificial (SIA) en educaci&#x00F3;n; y (ii) esbozar el nuevo perfil profesional docente para identificar las necesidades de formaci&#x00F3;n del profesorado.</p>
</sec>
<sec id="sec-2-31821-ES">
<label><bold>2.</bold></label>
<title><bold>C<sc>iencias de dise&#x00F1;o</sc>: <sc>objetivos</sc>, <sc>procesos y resultados</sc></bold></title>
<p>La Inteligencia Artificial en Educaci&#x00F3;n (AIED, por sus siglas en ingl&#x00E9;s) es un quehacer cient&#x00ED;fico en el &#x00E1;mbito de la ciencia de dise&#x00F1;o (<xref ref-type="bibr" rid="ref-23-31821-ES">Gonz&#x00E1;lez, 2017a</xref>). Su objetivo es crear herramientas inteligentes que permiten ampliar las posibilidades de diferentes usuarios en contextos espec&#x00ED;ficos de uso educativo. Surge de la confluencia de la Inteligencia Artificial y la Educaci&#x00F3;n, que son tambi&#x00E9;n disciplinas cient&#x00ED;ficas de Dise&#x00F1;o, pues buscan nuevas metas para ampliar las posibilidades humanas y se articulan en t&#x00E9;rminos de objetivos, procesos y resultados.</p>
<p>A diferencia de las Ciencias de la Naturaleza y las Ciencias Sociales, las Ciencias de Dise&#x00F1;o se orientan expresamente a predecir c&#x00F3;mo pueden llegar a ser las cosas y a prescribir pautas de actuaci&#x00F3;n que contribuyan a cambiar la realidad existente para que sea &#x201C;como debe ser&#x201D; (<xref ref-type="bibr" rid="ref-42-31821-ES">Simon, 1996</xref>; <xref ref-type="bibr" rid="ref-22-31821-ES">Gonz&#x00E1;lez, 2007</xref>). As&#x00ED;, su objetivo principal es modular el futuro de acuerdo con unos fines previamente elegidos.</p>
<p>Sobre esta base cient&#x00ED;fica se pueden dise&#x00F1;ar intervenciones en distintos campos de actividad humana y, para hacerlas operativas, pueden incorporar dise&#x00F1;os tecnol&#x00F3;gicos. Es el caso de los desarrollos de IA para resolver problemas y mejorar las posibilidades de la educaci&#x00F3;n en el logro de objetivos. Pero esto tiene profundas implicaciones &#x00E9;ticas que han de ser consideradas en relaci&#x00F3;n a varios aspectos:</p>
<p>(i) Los objetivos, que pueden ser discutibles o plantear conflictos entre las distintas partes interesadas. Ello incluye la selecci&#x00F3;n de los problemas cuya resoluci&#x00F3;n encomendamos a la IA. (ii) Los procesos: qu&#x00E9; procesos queremos automatizar, acelerar, etc. y qu&#x00E9; tecnolog&#x00ED;as de IA son adecuadas a ese prop&#x00F3;sito. (iii) Los resultados, que pueden conducir a cuestionar la intervenci&#x00F3;n misma (es decir, toda la secuencia de actividad pedag&#x00F3;gica mejorada con IA). (iv) Las consecuencias que se siguen a partir de las acciones realizadas, que pueden no ser las deseadas.</p>
</sec>
<sec id="sec-3-31821-ES">
<label><bold>3.</bold></label>
<title><bold>U<sc>sos y oportunidades de la ia en educaci&#x00F3;n</sc></bold></title>
<p>Entre la amplia diversidad de usos de la IA en la Educaci&#x00F3;n (<xref ref-type="bibr" rid="ref-17-31821-ES">Flores Vivar y Garc&#x00ED;a Pe&#x00F1;alvo, 2023</xref>; <xref ref-type="bibr" rid="ref-46-31821-ES">Uni&#x00F3;n Europea, 2022</xref>; <xref ref-type="bibr" rid="ref-30-31821-ES">Jara y Ochoa, 2020</xref>; <xref ref-type="bibr" rid="ref-38-31821-ES">Oca&#x00F1;a Fern&#x00E1;ndez <italic>et al</italic>., 2019</xref>; <xref ref-type="bibr" rid="ref-35-31821-ES">Moreno Padilla, 2019</xref>; <xref ref-type="bibr" rid="ref-41-31821-ES">S&#x00E1;nchez Vila y Lema Pen&#x00ED;n, 2007</xref>), destacan las siguientes:</p>
<list list-type="simple">
<list-item><label>(i)</label> <p>Facilitar la gesti&#x00F3;n, con la programaci&#x00F3;n de horarios, asignaci&#x00F3;n de recursos, etc.</p></list-item>
<list-item><label>(ii)</label> <p>Automatizar tareas rutinarias de los profesores, como el seguimiento de estudiantes, la provisi&#x00F3;n de informaci&#x00F3;n sobre los estudiantes a sus familias o tutores, la calificaci&#x00F3;n de ejercicios o pruebas, etc.</p></list-item>
<list-item><label>(iii)</label> <p>Apoyar la ense&#x00F1;anza, a trav&#x00E9;s de sistemas de tutor&#x00ED;a inteligentes que ofrecen asistencia a los estudiantes, en funci&#x00F3;n de las dificultades que pueda mostrar. Pueden ayudarles de manera efectiva, tambi&#x00E9;n fuera del aula, con la preparaci&#x00F3;n de las tareas. En contextos y situaciones en los que el docente humano no est&#x00E9; disponible, podr&#x00ED;an sustituirlo <italic>facilitadores virtuales</italic> (personajes virtuales realistas mediante la combinaci&#x00F3;n de tecnolog&#x00ED;as de IA, juegos en 3-D y animaci&#x00F3;n por ordenador).</p></list-item>
<list-item><label>(iv)</label> <p>Personalizar la experiencia educativa, que es quiz&#x00E1; la principal contribuci&#x00F3;n de la IA. Los llamados <italic>sistemas de ense&#x00F1;anza adaptativos</italic> son aplicaciones relacionadas con este objetivo. Sit&#x00FA;an a los estudiantes realmente en el centro, ajustando las trayectorias educativas a sus perfiles, caracter&#x00ED;sticas y comportamiento individual. Analizando la informaci&#x00F3;n obtenida acerca de su progreso, predicen el rendimiento futuro, lo que permite optimizar los recursos y el contenido que se les ofrece e incluso anticipar medidas correctoras para mejorarlo.</p></list-item>
</list>
<p>Contribuyen tambi&#x00E9;n a una educaci&#x00F3;n m&#x00E1;s personalizada distintas soluciones basadas en IA, creadas para atender a necesidades educativas especiales, como el subtitulado en directo en caso de discapacidad auditiva o la audiodescripci&#x00F3;n para quienes presentan discapacidad visual.</p>
<p>Cabe considerar, adem&#x00E1;s, otras muchas aplicaciones algor&#x00ED;tmicas que se incorporan al ecosistema de aprendizaje desde la educaci&#x00F3;n informal: a trav&#x00E9;s de las redes sociales, plataformas de juegos <italic>on-line</italic> o aplicaciones m&#x00F3;viles, entre otras.</p>
<p>Pero, aunque el uso de la IA en la Educaci&#x00F3;n se est&#x00E1; normalizando, la investigaci&#x00F3;n emp&#x00ED;rica sobre su efecto real sigue siendo escasa. Algunos estudios revelan que estas herramientas, con un enfoque basado en pasos para problemas bien definidos, est&#x00E1;n teniendo &#x00E9;xito en apoyar la ense&#x00F1;anza y el aprendizaje de conocimientos del dominio STEM (<xref ref-type="bibr" rid="ref-29-31821-ES">Humble &#x0026; Mozelius, 2022</xref>). Se trata, precisamente, de aquellas disciplinas que buscan desarrollar competencias aptas desde el punto de vista tecnol&#x00F3;gico y cient&#x00ED;fico para los retos del siglo XXI (<xref ref-type="bibr" rid="ref-35-31821-ES">Moreno Padilla, 2019</xref>). Tambi&#x00E9;n funcionan bien en el aprendizaje de lenguas extranjeras. En cualquier caso, parece conveniente mantener una actitud cr&#x00ED;tica y supervisada (<xref ref-type="bibr" rid="ref-46-31821-ES">Uni&#x00F3;n Europea, 2022</xref>, p. 14).</p>
</sec>
<sec id="sec-4-31821-ES">
<label><bold>4.</bold></label>
<title><bold>R<sc>iesgos potenciales e implicaciones &#x00E9;ticas</sc></bold></title>
<p>Para una adecuada implementaci&#x00F3;n de la IA en la educaci&#x00F3;n, preocupa la limitada formaci&#x00F3;n de los docentes, porque puede conducir al mal uso o incluso al abuso de esta tecnolog&#x00ED;a. Pero para comprender los riesgos potenciales, conviene mirar sobre todo a los dise&#x00F1;os de las herramientas y preguntarse a qu&#x00E9; responde su introducci&#x00F3;n en el &#x201C;mercado&#x201D; educativo, qui&#x00E9;n la impulsa y por qu&#x00E9; lo hace.</p>
<p>Los desarrollos de IA en educaci&#x00F3;n pueden interferir en la autonom&#x00ED;a y la responsabilidad de las personas y obstaculizar derechos universales (<xref ref-type="bibr" rid="ref-5-31821-ES">Asamblea General de la ONU, 1948</xref>) como la <italic>privacidad</italic> (art. 12), la <italic>igualdad</italic> (art. 1) <italic>y la no discriminaci&#x00F3;n</italic> (art. 2). Esto tiene implicaciones sociales y &#x00E9;ticas de car&#x00E1;cter general (<xref ref-type="bibr" rid="ref-14-31821-ES">Crawford, 2023</xref>) que se plantean tambi&#x00E9;n en otros &#x00E1;mbitos (<xref ref-type="bibr" rid="ref-31-31821-ES">Linares Salgado, 2022</xref>), pero requieren una especial precauci&#x00F3;n en contextos educativos.</p>
<p><italic>Privacidad</italic>: Los sistemas de aprendizaje autom&#x00E1;tico se entrenan con grandes vol&#x00FA;menes de datos. En educaci&#x00F3;n, se refieren a: (i) informaci&#x00F3;n personal sobre los estudiantes y sus familias; (ii) datos registrados acerca del rendimiento escolar; y (iii) datos de rastreo generados en el uso digital y las actividades de aprendizaje.</p>
<p>Los datos son la clave del &#x00E9;xito de la IA y constituyen la base de una educaci&#x00F3;n personalizada. Pero la l&#x00F3;gica del sector tecnol&#x00F3;gico seg&#x00FA;n la cual cualquier cosa puede ser un dato que se puede tomar y usar (<xref ref-type="bibr" rid="ref-14-31821-ES">Crawford, 2023</xref>) plantea, en el espacio educativo, una serie de problemas relacionados con: el modo c&#x00F3;mo se obtienen esos datos (el consentimiento y la privacidad); c&#x00F3;mo se analizan (la transparencia y la confianza); y la posibilidad de que puedan ser usados para fines distintos a los aprobados, de modo que los alumnos y sus familias acaben siendo v&#x00ED;ctimas de manipulaci&#x00F3;n comercial o de otro tipo (<xref ref-type="bibr" rid="ref-44-31821-ES">Unesco, 2019</xref>).</p>
<p>El riesgo de ciberataques es alto, cuando no existen protocolos de seguridad. Esto es especialmente preocupante cuando se trata de menores de edad. Los centros educativos est&#x00E1;n obligados a contar con procedimientos adecuados para garantizar la protecci&#x00F3;n y el uso &#x00E9;tico de los datos personales (<xref ref-type="bibr" rid="ref-46-31821-ES">UE, 2022</xref>, p. 11), pero la tarea no es f&#x00E1;cil cuando las habilidades y el conocimiento para desarrollar sistemas AIED se encuentra en organizaciones con fines de lucro y no dentro del sector educativo (<xref ref-type="bibr" rid="ref-29-31821-ES">Humble &#x0026; Mozelius, 2022</xref>).</p>
<p><italic>Igualdad y no discriminaci&#x00F3;n</italic>: Los datos con los que se entrenan los algoritmos de aprendizaje autom&#x00E1;tico incorporan sesgos provenientes de ciertos contextos y personas. Esto conduce a que estos sistemas internalicen criterios parciales o discriminatorios (de &#x201C;raza&#x201D;, g&#x00E9;nero, edad&#x2026;) propios de esas fuentes e incluso los amplifiquen.</p>
<p>Hist&#x00F3;ricamente, las distintas formas de desigualdad condicionaron el acceso a los recursos y las oportunidades y esto ahora incide en los datos con los que se entrenan los algoritmos para clasificar y reconocer patrones. As&#x00ED;, &#x201C;bajo el disfraz de una neutralidad t&#x00E9;cnica&#x201D; (<xref ref-type="bibr" rid="ref-14-31821-ES">Crawford, 2023</xref>, p. 200), que a menudo deja atr&#x00E1;s a los grupos de usuarios de los que se tienen menos datos disponibles., los sistemas de IA perpet&#x00FA;an las desigualdades sociales. El uso de la IA, tambi&#x00E9;n est&#x00E1; dilatando, socialmente, la brecha digital, ya que los pa&#x00ED;ses y los grupos humanos con m&#x00E1;s recursos son quienes se benefician de las oportunidades que ofrece. Todo ello compromete la equidad educativa.</p>
<p>Los sesgos son tambi&#x00E9;n consustanciales a las pr&#x00E1;cticas de clasificaci&#x00F3;n de la IA. Los dise&#x00F1;adores quienes tienen la &#x00FA;ltima palabra para decidir cu&#x00E1;les son las variables a considerar en el conjunto de datos de entrenamiento y qu&#x00E9; diferencias se tienen en cuenta para <italic>clasificar</italic> correctamente nuevas observaciones (<xref ref-type="bibr" rid="ref-14-31821-ES">Crawford, 2023</xref>). Esto contribuye a mantener y amplificar estereotipos (con claras implicaciones epistemol&#x00F3;gicas que trascienden el objetivo de este art&#x00ED;culo).</p>
<p><italic>Autonom&#x00ED;a:</italic> La interacci&#x00F3;n con los SIA puede obstaculizar el desarrollo de la autonom&#x00ED;a de los estudiantes y afectar a su capacidad de razonamiento y decisi&#x00F3;n, lo que plantea pues una dificultad a&#x00F1;adida para alcanzar el objetivo de convertirse en aprendices independientes (<xref ref-type="bibr" rid="ref-29-31821-ES">Humble &#x0026; Mozelius, 2022</xref>). Adem&#x00E1;s, renunciar a la libertad de elecci&#x00F3;n y delegar la toma de decisiones, socava la autonom&#x00ED;a como capacidad de autolegislarse y autodeterminarse, ligada al reconocimiento de la dignidad de las personas.</p>
<p><italic>Responsabilidad</italic>: Los SIA son agentes &#x2013; y pueden estar reemplazando la agencia humana&#x2013;, pero no son agentes morales. Esto genera problemas &#x00E9;ticos relacionados con la responsabilidad (e imputabilidad) de sus acciones, incluidas las responsabilidades legales que habr&#x00ED;a que afrontar en caso de que tuvieran consecuencias negativas para otras personas. Estos problemas se incrementan porque actualmente los algoritmos de aprendizaje autom&#x00E1;tico presentan elementos no predecibles debido a la opacidad con la que funcionan.</p>
<sec id="sec-5-31821-ES">
<label>4.1.</label>
<title><italic>Desaf&#x00ED;os &#x00E9;ticos espec&#x00ED;ficos del dominio educativo</italic></title>
<p>Junto a estas cuestiones generales, cada uno de los dominios en los que se aplica la IA presenta desaf&#x00ED;os &#x00E9;ticos espec&#x00ED;ficos que deben ser analizados en su contexto. En educaci&#x00F3;n, hay que tener en cuenta lo siguiente: 1) Los posibles da&#x00F1;os derivados de los diagn&#x00F3;sticos y la predicci&#x00F3;n de los resultados de aprendizaje de los estudiantes que pueden condicionar su desarrollo futuro. 2) Los problemas generados por las decisiones que podr&#x00ED;a tomar el sistema de IA por su influencia en las decisiones educativas de los docentes, las familias y otras partes interesadas (incluidos los legisladores). 3) Los impactos en el desarrollo y madurez de las personas, especialmente en las primeras etapas educativas, como consecuencia del cambio de roles que modifica la relaci&#x00F3;n entre docentes y discentes.</p>
<p>Tambi&#x00E9;n hay que considerar: 4) Las pedagog&#x00ED;as que orientan los dise&#x00F1;os de los sistemas de IA. Generalmente buscan aquello que se monetiza con mayor facilidad y son poco innovadoras. Contribuyen as&#x00ED; a mantener el <italic>statu quo</italic> del sistema educativo (<xref ref-type="bibr" rid="ref-28-31821-ES">Holmes <italic>et al</italic>, 2022</xref>), perdiendo su inmenso potencial para cambiar la educaci&#x00F3;n y ampliar realmente las oportunidades de las personas. 5) La imagen del mundo y las concepciones sociales y pol&#x00ED;ticas t&#x00E1;citas a las que responde la IA, que se trasladan al &#x00E1;mbito educativo. Esta tecnolog&#x00ED;a no es inocua y &#x201C;las clasificaciones que se eligen casualmente para darle forma a un sistema t&#x00E9;cnico pueden jugar un papel din&#x00E1;mico a la hora de delinear el mundo social y material&#x201D; (<xref ref-type="bibr" rid="ref-14-31821-ES">Crawford, 2023</xref>, p. 195). 6) Cabe citar tambi&#x00E9;n la dimensi&#x00F3;n &#x00E9;tica de algunas malas pr&#x00E1;cticas de los estudiantes sirvi&#x00E9;ndose de instrumentos de IA, que conllevan el enga&#x00F1;o y el fraude, atentan contra la propiedad y comprometen la honestidad intelectual.</p>
</sec>
</sec>
<sec id="sec-6-31821-ES">
<label><bold>5.</bold></label>
<title><bold>L<sc>a importancia de la &#x00E9;tica</sc></bold></title>
<p>Los aspectos positivos y los riesgos asociados a la IA est&#x00E1;n &#x00ED;ntimamente relacionados y no es posible acceder a las enormes ventajas que ofrece sin enfrentarnos a sus consecuencias negativas. Se hace necesario analizar y comprender estas ambivalencias para (i) anticipar y calibrar los impactos de determinados desarrollos y (ii) seleccionar los valores y principios que debemos proteger. Por eso ha surgido una &#x00E9;tica de la IA como &#x00E1;rea de investigaci&#x00F3;n multidisciplinaria en cuestiones de &#x00E9;tica aplicada relacionadas con los problemas normativos que plantea el desarrollo, despliegue y utilizaci&#x00F3;n de sistemas inteligentes. No es una &#x00E9;tica para sistemas inteligentes, sino para personas que dise&#x00F1;an, desarrollan o usan esos sistemas inteligentes, que son quienes pueden causar, aunque sea de manera involuntaria, los problemas morales.</p>
<p>El an&#x00E1;lisis normativo de los impactos de la IA, presenta paralelismos con otros subdominios espec&#x00ED;ficos en la &#x00E9;tica de la tecnolog&#x00ED;a (de internet, de datos, de robots&#x2026;) conformados en torno a distintas innovaciones. Con frecuencia abordan las mismas cuestiones (privacidad, sesgos&#x2026;) con planteamientos similares, pero se desarrollan sin interacci&#x00F3;n entre ellas y no llegan a conectar sus discusiones ni a ponerlas en relaci&#x00F3;n con los enfoques hist&#x00F3;ricos (<xref ref-type="bibr" rid="ref-40-31821-ES">S&#x00E6;tra &#x0026; Danaher, 2022</xref>). Esto conduce a duplicar esfuerzos sin aumentar el conocimiento real e incluso a obviar ideas consolidadas en la tradici&#x00F3;n &#x00E9;tica. As&#x00ED;, muchas veces, las dificultades para afrontar los desaf&#x00ED;os &#x00E9;ticos de la IA, proceden m&#x00E1;s del olvido de cuestiones fundamentales de la &#x00E9;tica &#x2013;y del conocimiento hist&#x00F3;ricamente adquirido&#x2013;, que de nuestra comprensi&#x00F3;n de la IA.</p>
<p>La &#x00E9;tica de la IA ha recibido mucha atenci&#x00F3;n por parte de los investigadores (<xref ref-type="bibr" rid="ref-8-31821-ES">Boddington, 2017</xref>; <xref ref-type="bibr" rid="ref-11-31821-ES">Coeckelberg, 2021</xref>; <xref ref-type="bibr" rid="ref-19-31821-ES">Floridi, 2022</xref>) y tambi&#x00E9;n de diversos organismos, instituciones y corporaciones. Actualmente, est&#x00E1; centrada en la elaboraci&#x00F3;n de c&#x00F3;digos &#x00E9;ticos para identificar los valores y principios que se deben poner en pr&#x00E1;ctica. Son m&#x00FA;ltiples las iniciativas y propuestas que surgieron en los &#x00FA;ltimos a&#x00F1;os y distintos pa&#x00ED;ses han intentado tambi&#x00E9;n establecer sugerencias y regulaciones. Con frecuencia, la proliferaci&#x00F3;n de c&#x00F3;digos con sus pautas y principios es vista como un problema (<xref ref-type="bibr" rid="ref-40-31821-ES">S&#x00E6;tra &#x0026; Danaher, 2022</xref>) ya que puede confundir (<xref ref-type="bibr" rid="ref-20-31821-ES">Floridi &#x0026; Cowls, 2021</xref>). Sin embargo, muchos de estos c&#x00F3;digos contemplan principios comunes, con ciertos &#x00E9;nfasis y matices afectados por valores culturales y sociales diferentes.</p>
<p>El an&#x00E1;lisis comparativo de seis c&#x00F3;digos de alto perfil para la IA &#x2013;publicados entre 2017 y 2018<xref ref-type="fn" rid="fn1-ES"><sup>1</sup></xref>&#x2013;, ha permitido a <xref ref-type="bibr" rid="ref-20-31821-ES">Floridi &#x0026; Cowls (2021)</xref> establecer un marco general con cinco principios b&#x00E1;sicos, cuatro de los cuales convergen con los utilizados en Bio&#x00E9;tica: <italic>beneficencia, no maleficencia, autonom&#x00ED;a</italic> y <italic>justicia</italic> (<xref ref-type="bibr" rid="ref-7-31821-ES">Beauchamp &#x0026; Childress 2019</xref>). No obstante, la transposici&#x00F3;n a los desaf&#x00ED;os de la IA no es perfecta y requiere, en algunos casos, una traducci&#x00F3;n adecuada:</p>
<p><italic>Beneficencia</italic> supone promover el bienestar, preservar la dignidad y el sostenimiento del planeta. <italic>No maleficencia</italic> remite a la privacidad, seguridad y precauci&#x00F3;n. <italic>Autonom&#x00ED;a</italic> se entiende como el poder de decidir. <italic>Justicia</italic> implica promover la prosperidad, preservar la solidaridad y evitar la injusticia. Cuestiones derivadas del desarrollo mismo de la IA hacen necesaria la incorporaci&#x00F3;n de un nuevo principio: la <italic>explicabilidad</italic>, que permite habilitar los otros principios a trav&#x00E9;s de la inteligibilidad y la rendici&#x00F3;n de cuentas (<xref ref-type="bibr" rid="ref-20-31821-ES">Floridi &#x0026; Cowls, 2021</xref>).</p>
<p>Este marco se presenta como la arquitectura dentro de la cual se puedan desarrollar las regulaciones, los est&#x00E1;ndares t&#x00E9;cnicos y las mejores pr&#x00E1;cticas para sectores espec&#x00ED;ficos. Puede desempe&#x00F1;ar tanto un papel habilitador (p. ej. el uso de la IA para alcanzar los Objetivos de Desarrollo Sostenible de las Naciones Unidas), como restrictivo (<xref ref-type="bibr" rid="ref-20-31821-ES">Floridi &#x0026; Cowls, 2021</xref>, p. 14). Es un marco &#x00E9;tico general y realista para el seguimiento y evaluaci&#x00F3;n del dise&#x00F1;o, desarrollo y uso de los SIA, porque dispone de un n&#x00FA;mero manejable de principios compatibles con valores fundamentales universalizables e irrenunciables. Estos principios ofrecen una base para las regulaciones espec&#x00ED;ficas que se propongan en una amplia variedad de contextos, para las leyes y acuerdos internacionales, las instancias de supervisi&#x00F3;n y evaluaci&#x00F3;n que los estados establezcan e incluso los observatorios ciudadanos&#x2026; Proporcionan, de ese modo, legitimidad moral, jur&#x00ED;dica y pol&#x00ED;tica.</p>
</sec>
<sec id="sec-7-31821-ES">
<label><bold>6.</bold></label>
<title><bold>U<sc>n marco &#x00E9;tico unificado para la inteligencia artificial en el contexto educativo</sc></bold></title>
<p>Desde las ideas expuestas, se puede avanzar de lo general a lo particular. Esto requiere adoptar marcos &#x00E9;ticos para el uso y desarrollo de la IA en la educaci&#x00F3;n. Su importancia ha sido se&#x00F1;alada en investigaciones anteriores (<xref ref-type="bibr" rid="ref-28-31821-ES">Holmes <italic>et al</italic>., 2022</xref>; <xref ref-type="bibr" rid="ref-27-31821-ES">Holmes y Porayska-Pomsta, 2023</xref>). Pero desde las D<italic>irectrices para una &#x00C9;tica de la IA en Educaci&#x00F3;n</italic> que Aiken y Epstein publicaron hace m&#x00E1;s de dos d&#x00E9;cadas (<xref ref-type="bibr" rid="ref-2-31821-ES">Aiken &#x0026; Epstein 2000</xref>), ha habido una sorprendente escasez de trabajos publicados que se centren expl&#x00ED;citamente en la &#x00E9;tica (<xref ref-type="bibr" rid="ref-28-31821-ES">Holmes, <italic>et al.</italic>, 2022</xref>, pp. 505-506). Recientemente, esta cuesti&#x00F3;n volvi&#x00F3; a suscitar el inter&#x00E9;s de la comunidad investigadora en educaci&#x00F3;n (<xref ref-type="bibr" rid="ref-10-31821-ES">Chu <italic>et al</italic>, 2022</xref>; <xref ref-type="bibr" rid="ref-27-31821-ES">Holmes y Porayska-Pomsta, 2023</xref>; <xref ref-type="bibr" rid="ref-9-31821-ES">Boddington, 2023</xref>; <xref ref-type="bibr" rid="ref-36-31821-ES">Nguyen <italic>et al.</italic>, 2023</xref>) y, en los &#x00FA;ltimos a&#x00F1;os, surgieron tambi&#x00E9;n algunas iniciativas relevantes de organizaciones internacionales (<xref ref-type="bibr" rid="ref-44-31821-ES">Unesco, 2019</xref>; <xref ref-type="bibr" rid="ref-46-31821-ES">Uni&#x00F3;n Europea, 2022</xref>).</p>
<p>La Uni&#x00F3;n Europea, dentro de su Plan de Acci&#x00F3;n para la Educaci&#x00F3;n Digital (2021-2027) -cuyo fin es apoyar una adaptaci&#x00F3;n sostenible y eficaz de los sistemas de educaci&#x00F3;n y formaci&#x00F3;n de los Estados miembros a la era digital- estableci&#x00F3; una serie de <italic>directrices &#x00E9;ticas sobre el uso de la Inteligencia Artificial (IA) y los datos en la educaci&#x00F3;n y formaci&#x00F3;n para los educadores</italic> (<xref ref-type="bibr" rid="ref-46-31821-ES">Uni&#x00F3;n Europea, 2022</xref>). En esas directrices, basadas en los principios y requisitos para una IA fiable (<xref ref-type="bibr" rid="ref-12-31821-ES">Uni&#x00F3;n Europea, 2019</xref>), hay cuatro consideraciones clave que deben orientan a los agentes educativos en sus decisiones sobre el uso de la IA: la capacidad de acci&#x00F3;n humana, la equidad, la humanidad y la elecci&#x00F3;n justificada.</p>
<p>Dentro de la comunidad acad&#x00E9;mica, se han analizado de manera sistem&#x00E1;tica directrices e informes &#x00E9;ticos de organizaciones internacionales y se han seleccionado principios aplicables para garantizar que los sistemas de IA desarrollados para la educaci&#x00F3;n, sean esencialmente &#x00E9;ticos por dise&#x00F1;o: Principio de gobernanza y administraci&#x00F3;n, Principio de transparencia y responsabilidad, Principio de sostenibilidad y proporcionalidad, Principio de privacidad, Principio de seguridad y protecci&#x00F3;n, Principio de inclusi&#x00F3;n y Principio de AIED centrado en el ser humano (<xref ref-type="bibr" rid="ref-36-31821-ES">Nguyen <italic>et al</italic>. 2023</xref>).</p>
<p>Consideradas estas propuestas, hace falta contar con un marco unificado para una IA responsable, &#x00E9;tica y confiable. Con ello se busca evitar la superposici&#x00F3;n de principios y recomendaciones para cada dominio espec&#x00ED;fico. Existe un marco de prestigio, que ha desempe&#x00F1;ado un importante papel en los debates del Grupo de expertos de alto nivel en IA de la Comisi&#x00F3;n Europea e influido en distintas recomendaciones (<xref ref-type="bibr" rid="ref-20-31821-ES">Floridi &#x0026; Colws, 2021</xref>). Sus principios se adaptan sorprendentemente bien a una amplia gama de contextos (<xref ref-type="bibr" rid="ref-31-31821-ES">Linares Salgado, 2022</xref>). La cuesti&#x00F3;n es si pueden regular tambi&#x00E9;n problemas relativos al cu&#x00E1;ndo, c&#x00F3;mo y con qu&#x00E9; fin se utiliza en educaci&#x00F3;n un sistema inteligente, para situar decisiones, acciones y pr&#x00E1;cticas pedag&#x00F3;gicas concretas que aseguren un uso adecuado de la IA.</p>
<p>Esta cuesti&#x00F3;n llev&#x00F3; a un nuevo an&#x00E1;lisis en tres fases:</p>
<list list-type="simple">
<list-item><label>1)</label> <p>Revisi&#x00F3;n de gu&#x00ED;as y recomendaciones que contienen c&#x00F3;digos &#x00E9;ticos, relevantes por su impacto, publicados posteriormente a los examinados por <xref ref-type="bibr" rid="ref-20-31821-ES">Floridi &#x0026; Cowls (2021)</xref>. El objetivo es comprobar la vigencia y prevalencia de los principios &#x00E9;ticos que proponen. En el muestro se busc&#x00F3; que estuvieran representadas las diversas perspectivas influyentes del sector: a) recomendaciones de Organizaciones internacionales: <xref ref-type="bibr" rid="ref-12-31821-ES">UE (2019)</xref>, <xref ref-type="bibr" rid="ref-45-31821-ES">UNESCO (2022)</xref> y <xref ref-type="bibr" rid="ref-39-31821-ES">OCDE (2019)</xref>; b) documentos de pa&#x00ED;ses implicados en la carrera por la IA: EE UU (<xref ref-type="bibr" rid="ref-16-31821-ES">Executive Office of the President, 2020</xref>), China (<xref ref-type="bibr" rid="ref-34-31821-ES">Ministerio de Ciencia y Tecnolog&#x00ED;a de la Rep&#x00FA;blica Popular China, 2017</xref>), <xref ref-type="bibr" rid="ref-12-31821-ES">UE (2019)</xref> y Espa&#x00F1;a (<xref ref-type="bibr" rid="ref-21-31821-ES">Gobierno de Espa&#x00F1;a, 2020</xref>); y c) iniciativas de varias empresas relevantes en el sector: <xref ref-type="bibr" rid="ref-33-31821-ES">Microsoft (2022)</xref>, <xref ref-type="bibr" rid="ref-26-31821-ES">Google (2022)</xref> y Meta (<xref ref-type="bibr" rid="ref-32-31821-ES">Meta Platforms, 2021</xref>).</p></list-item>
<list-item><label>2)</label> <p>Comprobaci&#x00F3;n de la convergencia, en el marco &#x00E9;tico anterior, de: (i) los principios seleccionados por <xref ref-type="bibr" rid="ref-36-31821-ES">Nguyen <italic>et al</italic>. (2023)</xref> y (ii) las directrices &#x00E9;ticas de la Comisi&#x00F3;n Europea para guiar el uso de la Inteligencia Artificial en educaci&#x00F3;n.</p></list-item>
<list-item><label>3)</label> <p>Constataci&#x00F3;n de que los valores que orientan estos principios est&#x00E1;n alineados con los valores propios de la educaci&#x00F3;n (<xref ref-type="bibr" rid="ref-3-31821-ES">Alonso, 2022</xref> y <xref ref-type="bibr" rid="ref-4-31821-ES">2023</xref>).</p></list-item>
</list>
<sec id="sec-8-31821-ES">
<label>6.1.</label>
<title><italic>Resultados</italic></title>
<p>El an&#x00E1;lisis comparativo, como puede verse en la <xref ref-type="table" rid="tabw-1-31821-ES">Tabla 1</xref>, muestra un alto grado de superposici&#x00F3;n entre los principios de los nueve c&#x00F3;digos, que convergen claramente en el marco general propuesto por <xref ref-type="bibr" rid="ref-20-31821-ES">Floridi &#x0026; Cowls (2021)</xref>. Aunque solo la <italic>explicabilidad</italic> aparece con id&#x00E9;ntica formulaci&#x00F3;n &#x2013;o como <italic>transparencia</italic>&#x2013; en la totalidad de los c&#x00F3;digos, la &#x201C;traducci&#x00F3;n&#x201D; realizada para transponer los de la bio&#x00E9;tica a la IA, permite identificarlos.</p>
<table-wrap id="tabw-1-31821-ES">
<label>T<sc>abla</sc> 1</label>
<caption><title>V<sc>isi&#x00F3;n sin&#x00F3;ptica</sc> 1</title></caption>
<table id="tab-1-31821-ES" frame="hsides" border="1" rules="all">
<col width="16%"/>
<col width="17%"/>
<col width="17%"/>
<col width="16%"/>
<col width="17%"/>
<col width="17%"/>
<thead>
<tr>
<th valign="top" align="left"><p>&#x00A0;</p></th>
<th valign="top" align="left"><p><bold>Beneficencia</bold></p></th>
<th valign="top" align="left"><p><bold>No Maledicencia</bold></p></th>
<th valign="top" align="left"><p><bold>Autonom&#x00ED;a</bold></p></th>
<th valign="top" align="left"><p><bold>Justicia</bold></p></th>
<th valign="top" align="left"><p><bold>Explicabilidad</bold></p></th>
</tr>
</thead>
<tbody>
<tr>
<td valign="top" align="left"><p>&#x00A0;</p></td>
<td valign="top" align="left"><p>Promover el bienestar, preservar la Dignidad y el Sostenimiento del Planeta</p></td>
<td valign="top" align="left"><p>Privacidad, Seguridad y &#x2018;Precauci&#x00F3;n de Capacidad&#x2019;</p></td>
<td valign="top" align="left"><p>Poder de decidir (Decidir)</p></td>
<td valign="top" align="left"><p>Promover la prosperidad, preservar la solidaridad, evitar la injusticia.</p></td>
<td valign="top" align="left"><p>Habilitar los otros principios a trav&#x00E9;s de la inteligibilidad y la rendici&#x00F3;n de cuentas</p></td>
</tr>
<tr>
<td valign="top" align="left"><p><bold>UNESCO</bold></p></td>
<td valign="top" align="left"><p>Sostenibilidad</p></td>
<td valign="top" align="left"><p>Proporcionalidad y no hacer da&#x00F1;o. Seguridad y protecci&#x00F3;n</p></td>
<td valign="top" align="left"><p>Supervisi&#x00F3;n y decisi&#x00F3;n humanas</p></td>
<td valign="top" align="left"><p>Equidad y no discriminaci&#x00F3;n</p></td>
<td valign="top" align="left"><p>Transparencia y explicabilidad</p></td>
</tr>
<tr>
<td valign="top" align="left"><p><bold>OCDE</bold></p></td>
<td valign="top" align="left"><p>Crecimiento inclusivo, desarrollo sostenible y bienestar</p></td>
<td valign="top" align="left"><p>Robustez, seguridad y protecci&#x00F3;n</p></td>
<td valign="top" align="left"><p>Respetar los valores humanos y la justicia (incluye la autonom&#x00ED;a)</p></td>
<td valign="top" align="left"><p>Respetar los valores humanos y la justicia</p></td>
<td valign="top" align="left"><p>Transparencia y explicabilidad Responsabilidad</p></td>
</tr>
<tr>
<td valign="top" align="left"><p><bold>UE</bold></p></td>
<td valign="top" align="left"><p>&#x00A0;</p></td>
<td valign="top" align="left"><p>Prevenci&#x00F3;n del da&#x00F1;o</p></td>
<td valign="top" align="left"><p>Respeto por la autonom&#x00ED;a humana</p></td>
<td valign="top" align="left"><p>Equidad</p></td>
<td valign="top" align="left"><p>Explicabilidad</p></td>
</tr>
<tr>
<td valign="top" align="left"><p><bold>EE. UU.</bold></p></td>
<td valign="top" align="left"><p>&#x00A0;</p></td>
<td valign="top" align="left"><p>Seguridad y protecci&#x00F3;n</p></td>
<td valign="top" align="left"><p>&#x00A0;</p></td>
<td valign="top" align="left"><p>Equidad y no discriminaci&#x00F3;n</p></td>
<td valign="top" align="left"><p>Divulgaci&#x00F3;n y transparencia</p></td>
</tr>
<tr>
<td valign="top" align="left"><p><bold>CHINA</bold></p></td>
<td valign="top" align="left"><p>Armon&#x00ED;a y amistad: promover el bienestar com&#x00FA;n</p></td>
<td valign="top" align="left"><p>Privacidad: respetar y proteger la privacidad</p></td>
<td valign="top" align="left"><p>Privacidad: respetar y proteger el derecho de las personas de conocer y elegir</p></td>
<td valign="top" align="left"><p>Equidad y Justicia</p></td>
<td valign="top" align="left"><p>Seguridad: transparencia, explicabilidad, trazabilidad, confiabilidad, auditabilidad y seguridad</p></td>
</tr>
<tr>
<td valign="top" align="left"><p><bold>ESPA&#x00D1;A</bold></p></td>
<td valign="top" align="left"><p>Bienestar social Sostenibilidad</p></td>
<td valign="top" align="left"><p>Gobierno de datos y sistemas: los datos no pueden usarse para perjudicar a las personas o violar sus derechos fundamentales</p></td>
<td valign="top" align="left"><p>Bienestar social: no reducir, limitar o desviar la autonom&#x00ED;a de las personas</p></td>
<td valign="top" align="left"><p>Inclusi&#x00F3;n</p></td>
<td valign="top" align="left"><p>Transparencia</p></td>
</tr>
<tr>
<td valign="top" align="left"><p><bold>GOOGLE</bold></p></td>
<td valign="top" align="left"><p>Beneficio a la sociedad</p></td>
<td valign="top" align="left"><p>Garantizar la seguridad de los sistemas inteligentes Incorporar principios de dise&#x00F1;o de privacidad</p></td>
<td valign="top" align="left"><p>Ser responsables ante las personas: los sistemas deben estar sujetos a un adecuado control y direcci&#x00F3;n humana</p></td>
<td valign="top" align="left"><p>Evitar crear y reforzar sesgos</p></td>
<td valign="top" align="left"><p>Ser responsables ante las personas: proveer oportunidades adecuadas de respuesta, explicaci&#x00F3;n y apelaci&#x00F3;n</p></td>
</tr>
<tr>
<td valign="top" align="left"><p><bold>MICRO-SOFT</bold></p></td>
<td valign="top" align="left"><p>&#x00A0;</p></td>
<td valign="top" align="left"><p>Privacidad y Seguridad</p></td>
<td valign="top" align="left"><p>&#x00A0;</p></td>
<td valign="top" align="left"><p>Equidad</p></td>
<td valign="top" align="left"><p>Transparencia Responsabilidad</p></td>
</tr>
<tr>
<td valign="top" align="left"><p><bold>META</bold></p></td>
<td valign="top" align="left"><p>&#x00A0;</p></td>
<td valign="top" align="left"><p>Solidez y seguridad</p></td>
<td valign="top" align="left"><p>&#x00A0;</p></td>
<td valign="top" align="left"><p>Equidad e inclusi&#x00F3;n</p></td>
<td valign="top" align="left"><p>Transparencia y control</p></td>
</tr>
</tbody>
</table>
<table-wrap-foot>
<attrib>Fuente: elaboraci&#x00F3;n propia</attrib>
</table-wrap-foot>
</table-wrap>
<list list-type="simple">
<list-item><label>I)</label> <p>El principio de <italic>no-maledicencia</italic> (que remite al <italic>non-laedere</italic> del derecho romano)<xref ref-type="fn" rid="fn2-ES"><sup>2</sup></xref> figura expresamente como <italic>prevenci&#x00F3;n del da&#x00F1;o</italic> en las directrices de la UE, como <italic>seguridad y protecci&#x00F3;n</italic> en seis c&#x00F3;digos y centrada en la protecci&#x00F3;n de la <italic>privacidad</italic> en otros dos. En la <italic>Estrategia Nacional para la IA</italic> espa&#x00F1;ola, se asocia la prevenci&#x00F3;n al principio de Gobierno de datos y sistemas: &#x201C;los datos no pueden usarse para perjudicar a las personas o violar sus derechos fundamentales&#x201D; (<xref ref-type="bibr" rid="ref-21-31821-ES">Gobierno de Espa&#x00F1;a, 2020</xref>, p. 67).</p></list-item>
<list-item><label>II)</label> <p>La <italic>justicia</italic> <bold>(</bold><italic>suum cuique tribuere</italic>) se cita de forma expl&#x00ED;cita en uno de los c&#x00F3;digos (China), siendo m&#x00E1;s frecuente la <italic>equidad</italic> como criterio de justicia (en siete de ellos). Es reconocible en la <italic>inclusi&#x00F3;n</italic> (Espa&#x00F1;a) y en la prevenci&#x00F3;n de amenazas para la justicia como <italic>evitar crear y reforzar sesgos</italic> (Google) y la <italic>no discriminaci&#x00F3;n.</italic></p></list-item>
<list-item><label>III)</label> <p>El principio de <italic>autonom&#x00ED;a</italic> est&#x00E1; asociado a la toma de decisiones dirigida por humanos, que puede ponerse en riesgo a medida que aumenta la agencia de las m&#x00E1;quinas. Es llamativo que este principio no est&#x00E9; recogido en la &#x201C;Gu&#x00ED;a para la regulaci&#x00F3;n de aplicaciones de IA&#x201D; de los EEUU y en dos de las tres iniciativas de la industria privada (Microsoft y Meta). Aparece en los dem&#x00E1;s c&#x00F3;digos bajo la formulaci&#x00F3;n expresa de <italic>autonom&#x00ED;a</italic>, como <italic>supervisi&#x00F3;n, control</italic> o <italic>direcci&#x00F3;n humanas</italic> o como <italic>derecho a decidir</italic>.</p></list-item>
<list-item><label>IV)</label> <p>Los mismos c&#x00F3;digos que no incluyen la autonom&#x00ED;a entre sus principios, tampoco incorporan el de <italic>beneficencia</italic>, que limita la creaci&#x00F3;n de SIA a aquellos que han de ser beneficiosos para la humanidad. Tampoco aparece expresamente en las directrices &#x00E9;ticas de la UE, sin embargo, el <italic>bienestar social y medioambiental</italic> s&#x00ED; figura entre los requisitos que derivan del principio de <italic>equidad</italic>. En el resto de los c&#x00F3;digos presenta formulaciones diversas, vinculado al <italic>bienestar</italic> y la <italic>sostenibilidad</italic>.</p></list-item>
</list>
<p>Cabe concluir que el marco presentado por <xref ref-type="bibr" rid="ref-20-31821-ES">Floridi &#x0026; Cowls (2021)</xref> ofrece una relaci&#x00F3;n suficiente y completa de los principios &#x00E9;ticos clave para la IA. Puede resultar &#x00FA;til &#x2013;como sugieren los autores&#x2013; para orientar regulaciones y pr&#x00E1;cticas en dominios espec&#x00ED;ficos. Deber&#x00ED;a recoger tambi&#x00E9;n los principios que se proponen en el &#x00E1;mbito educativo, encaminados en dos direcciones: (i) dirigir el desarrollo, despliegue y uso de las herramientas y (ii) orientar a los agentes educativos en sus pr&#x00E1;cticas profesionales.</p>
<p>Es claro el paralelismo entre los principios de <italic>no maleficencia</italic> y <italic>beneficencia</italic> y los &#x201C;metaprincipios&#x201D; que propusieron Aiken y Epstein como fundamento filos&#x00F3;fico b&#x00E1;sico para cualquier discusi&#x00F3;n sobre los sistemas AIED. Seg&#x00FA;n el <italic>meta-principio negativo</italic> <bold>&#x201C;</bold>la tecnolog&#x00ED;a AIED no debe disminuir al estudiante en ninguna de las dimensiones fundamentales del ser humano&#x201D;. El <italic>meta-principio positivo</italic> establece que <bold>&#x201C;</bold>la tecnolog&#x00ED;a AIED debe potenciar al estudiante en al menos una de las dimensiones fundamentales del ser humano&#x201D; (<xref ref-type="bibr" rid="ref-2-31821-ES">Aiken &#x0026; Epstein, 2000</xref>, p. 170).</p>
<p>Tambi&#x00E9;n, como se observa en la <xref ref-type="table" rid="tabw-2-31821-ES">Tabla 2</xref>, la propuesta de <xref ref-type="bibr" rid="ref-36-31821-ES">Nguyen y colaboradores (2023)</xref> presenta un alto grado de coherencia con el marco unificado: (i) El principio de <italic>no maleficencia</italic>, implica <italic>privacidad, seguridad</italic> y <italic>precauci&#x00F3;n de capacidad</italic>. Incorpora, por lo tanto, los <italic>principios de privacidad y de seguridad y protecci&#x00F3;n.</italic> (ii) <italic>Beneficencia</italic> supone <italic>promover el bienestar, preservar la dignidad</italic> y el <italic>sostenimiento del planeta</italic>, donde se reconoce el <italic>principio de sostenibilidad y proporcionalidad</italic>. (iii) <italic>Autonom&#x00ED;a</italic> significa <italic>poder de decidir</italic>. Ligada al reconocimiento de la dignidad de las personas conduce al principio b&#x00E1;sico que debe presidir nuestra relaci&#x00F3;n con la IA: la <italic>irremplazabilidad de lo humano,</italic> que exige el <italic>principio de IA educativa centrado en el ser humano.</italic> (iv) <italic>Justicia</italic> es <italic>promover la prosperidad, preservar la solidaridad, evitar la injusticia,</italic> lo que persigue el <italic>principio de inclusi&#x00F3;n.</italic> (v) La <italic>explicabilidad</italic> permite habilitar los otros principios a trav&#x00E9;s de la <italic>inteligibilidad</italic> y la <italic>rendici&#x00F3;n de cuentas</italic>. La inteligibilidad requiere <italic>transparencia</italic> y la <italic>rendici&#x00F3;n de cuentas</italic> supone atribuci&#x00F3;n de <italic>responsabilidades</italic>, que es lo que exige el <italic>principio de transparencia y responsabilidad.</italic></p>
<table-wrap id="tabw-2-31821-ES">
<label>T<sc>abla</sc> 2</label>
<caption><title>V<sc>isi&#x00F3;n sin&#x00F3;ptica</sc> 2</title></caption>
<table id="tab-2-31821-ES" frame="hsides" border="1" rules="all">
<col width="20%"/>
<col width="20%"/>
<col width="20%"/>
<col width="20%"/>
<col width="20%"/>
<thead>
<tr>
<th valign="top" align="left" rowspan="2" colspan="2"><p><bold><xref ref-type="bibr" rid="ref-20-31821-ES">Floridi &#x0026; Cowls, 2021</xref></bold></p></th>
<th valign="top" align="center" colspan="3"><p><bold>AIED</bold></p></th>
</tr>
<tr>
<th valign="top" align="left"><p><bold><xref ref-type="bibr" rid="ref-36-31821-ES">Nguyen <italic>et al</italic>., 2023</xref></bold></p></th>
<th valign="top" align="left" colspan="2"><p><bold>Valores de la Educaci&#x00F3;n (<xref ref-type="bibr" rid="ref-46-31821-ES">UE, 2022</xref>)</bold></p></th>
</tr>
<tr>
<th valign="top" align="left"><p><bold>Principios</bold></p></th>
<th valign="top" align="left"><p>&#x00A0;</p></th>
<th valign="top" align="left"><p><bold>Principios</bold></p></th>
<th valign="top" align="left"><p><bold>Consideraciones &#x00E9;ticas</bold></p></th>
<th valign="top" align="left"><p><bold>Valores</bold></p></th>
</tr>
</thead>
<tbody>
<tr>
<td valign="top" align="left" rowspan="3"><p><bold>No maleficencia:</bold></p></td>
<td valign="top" align="left"><p>Privacidad</p></td>
<td valign="top" align="left"><p>Privacidad</p></td>
<td valign="top" align="left" rowspan="3"><p>&#x00A0;</p></td>
<td valign="top" align="left" rowspan="3"><p>&#x00A0;</p></td>
</tr>
<tr>
<td valign="top" align="left"><p>Seguridad</p></td>
<td valign="top" align="left"><p>Seguridad y protecci&#x00F3;n</p></td>
</tr>
<tr>
<td valign="top" align="left"><p>&#x2018;Precauci&#x00F3;n de Capacidad&#x2019;</p></td>
<td valign="top" align="left"><p>&#x00A0;</p></td>
</tr>
<tr>
<td valign="top" align="left" rowspan="5"><p><bold>Beneficencia:</bold></p></td>
<td valign="top" align="left" rowspan="2"><p>Promover el Bienestar</p></td>
<td valign="top" align="left" rowspan="2"><p>&#x00A0;</p></td>
<td valign="top" align="left" rowspan="5"><p>La humanidad (consideraci&#x00F3;n para las personas, identidad y dignidad)</p></td>
<td valign="top" align="left"><p>Bienestar</p></td>
</tr>
<tr>
<td valign="top" align="left"><p>Seguridad</p></td>
</tr>
<tr>
<td valign="top" align="left" rowspan="2"><p>Preservar la Dignidad</p></td>
<td valign="top" align="left" rowspan="2"><p>&#x00A0;</p></td>
<td valign="top" align="left"><p>Cohesi&#x00F3;n social</p></td>
</tr>
<tr>
<td valign="top" align="left"><p>Contacto valioso</p></td>
</tr>
<tr>
<td valign="top" align="left"><p>Sostenimiento del Planeta</p></td>
<td valign="top" align="left"><p>Sostenibilidad y proporcionalidad</p></td>
<td valign="top" align="left"><p>Conexi&#x00F3;n humana significativa</p></td>
</tr>
<tr>
<td valign="top" align="left" rowspan="3"><p><bold>Autonom&#x00ED;a:</bold></p></td>
<td valign="top" align="left" rowspan="3"><p>Poder de decidir (decidir) (Conduce a la irreemplazabilidad de lo humano)</p></td>
<td valign="top" align="left" rowspan="3"><p>IA educativa centrada en el ser humano</p></td>
<td valign="top" align="left" rowspan="3"><p>La capacidad de acci&#x00F3;n humana</p></td>
<td valign="top" align="left"><p>Autonom&#x00ED;a</p></td>
</tr>
<tr>
<td valign="top" align="left"><p>Autodeterminaci&#x00F3;n</p></td>
</tr>
<tr>
<td valign="top" align="left"><p>Responsabilidad</p></td>
</tr>
<tr>
<td valign="top" align="left" rowspan="4"><p><bold>Justicia:</bold></p></td>
<td valign="top" align="left" rowspan="2"><p>Promover la prosperidad</p></td>
<td valign="top" align="left" rowspan="4"><p>Principio de inclusi&#x00F3;n</p></td>
<td valign="top" align="left" rowspan="4"><p>La equidad (trato justo)</p></td>
<td valign="top" align="left"><p>Igualdad de oportunidades</p></td>
</tr>
<tr>
<td valign="top" align="left"><p>Inclusi&#x00F3;n</p></td>
</tr>
<tr>
<td valign="top" align="left"><p>Preservar la solidaridad</p></td>
<td valign="top" align="left"><p>No discriminaci&#x00F3;n</p></td>
</tr>
<tr>
<td valign="top" align="left"><p>Evitar la injusticia</p></td>
<td valign="top" align="left"><p>Distribuci&#x00F3;n justa de derechos y responsabilidades</p></td>
</tr>
<tr>
<td valign="top" align="left" rowspan="2"><p><bold>Explicabilidad:</bold></p></td>
<td valign="top" align="left"><p>Inteligibilidad</p></td>
<td valign="top" align="left" rowspan="2"><p>Principio de transparencia y responsabilidad</p></td>
<td valign="top" align="left" rowspan="2"><p>La elecci&#x00F3;n justificada</p>
<p>(uso de conocimientos, hechos y datos para justificar las elecciones de las distintas partes interesadas en el entorno escolar)</p></td>
<td valign="top" align="left"><p>Transparencia</p></td>
</tr>
<tr>
<td valign="top" align="left"><p>Rendici&#x00F3;n de cuentas</p>
<p>(La inteligibilidad requiere transparencia y la rendici&#x00F3;n de cuentas supone atribuci&#x00F3;n de responsabilidad-des)</p></td>
<td valign="top" align="left"><p>Explicabilidad</p></td>
</tr>
</tbody>
</table>
<table-wrap-foot>
<attrib>Fuente: elaboraci&#x00F3;n propia</attrib>
</table-wrap-foot>
</table-wrap>
<p>Asimismo, est&#x00E1;n en sinton&#x00ED;a con estos principios las consideraciones &#x00E9;ticas identificadas por la Uni&#x00F3;n Europea para orientar a los agentes educativos en sus decisiones sobre la IA (<xref ref-type="bibr" rid="ref-46-31821-ES">Uni&#x00F3;n Europea, 2022</xref>, p. 18). Se reconocen en ellas: (i) los valores que constituyen los fines de la Educaci&#x00F3;n (autonom&#x00ED;a, autodeterminaci&#x00F3;n, responsabilidad, cohesi&#x00F3;n social, bienestar, etc.); (ii) los valores indispensables para su desarrollo (igualdad de oportunidades, no discriminaci&#x00F3;n, transparencia, explicabilidad&#x2026;); y (iii): los valores que sirven a la educaci&#x00F3;n como criterio de juicio, entre ellos la inclusi&#x00F3;n, la distribuci&#x00F3;n justa de derechos y responsabilidades, o una conexi&#x00F3;n humana significativa (<xref ref-type="bibr" rid="ref-3-31821-ES">Alonso, 2022</xref> y <xref ref-type="bibr" rid="ref-4-31821-ES">2023</xref>).</p>
<p>Los principios &#x00E9;ticos de <italic>beneficencia, no maleficencia, autonom&#x00ED;a, justicia</italic> y <italic>explicabilidad</italic> aplicados a los SIA, pueden prevenir los riesgos asociados a estos sistemas para usarlos en beneficio del Derecho a la Educaci&#x00F3;n (ODS 4), como fin en s&#x00ED; mismo y medio para alcanzar todos los dem&#x00E1;s derechos. Los instrumentos que &#x2014;en su dise&#x00F1;o, despliegue y uso&#x2014; pudieran conculcar alguno de estos principios, no ser&#x00ED;an leg&#x00ED;timos al poner en peligro el respeto a la dignidad humana que los derechos humanos preservan.</p>
<p><italic>No maleficencia.</italic> La IA debe dise&#x00F1;arse de modo que no agrave da&#x00F1;os existentes ni cree otros nuevos. As&#x00ED;, no ser&#x00ED;a &#x00E9;tico dise&#x00F1;ar instrumentos que obstaculicen el desarrollo de alguna de las capacidades propias de la inteligencia humana en sus tres dominios principales: la percepci&#x00F3;n &#x2014;conocimiento&#x2014;, la volici&#x00F3;n y los valores (<xref ref-type="bibr" rid="ref-24-31821-ES">Gonz&#x00E1;lez, 2017b</xref>), limitando el uso de los sentidos, la imaginaci&#x00F3;n, el pensamiento y el razonamiento de un modo &#x201C;verdaderamente humano&#x201D; (<xref ref-type="bibr" rid="ref-37-31821-ES">Nussbaum, 2012</xref>, p. 53).</p>
<p><italic>Beneficencia.</italic> Adem&#x00E1;s de no da&#x00F1;ar, los SIA educativos deben dise&#x00F1;arse para que los estudiantes puedan desarrollar su potencial y usar todas sus capacidades para elegir la vida que desean vivir de forma plena y creativa, conforme a la dignidad humana.</p>
<p><italic>Autonom&#x00ED;a.</italic> En el &#x00E1;mbito de la &#x00E9;tica se entiende por <italic>autonom&#x00ED;</italic>a la capacidad de autolegislarse y autodeterminarse, que solo es aplicable a los seres humanos. Esto impone y justifica el control humano sobre la &#x201C;m&#x00E1;quina&#x201D;. Por eso: (a) No pueden delegarse en m&#x00E1;quinas inteligentes, sin supervisi&#x00F3;n humana, decisiones que afectan a la vida de las personas. En consecuencia, las decisiones educativas no pueden estar basadas &#x00FA;nicamente en el proceso autom&#x00E1;tico de datos. (b) Es exigible que un humano sepa siempre si est&#x00E1; interactuando con otro humano o con una m&#x00E1;quina. As&#x00ED;, cuando se usen instrumentos de IA, las precauciones especiales con personas vulnerables, como los ni&#x00F1;os, hacen necesario el consentimiento informado de los padres o tutores legales.</p>
    <p>La educaci&#x00F3;n constituye un &#x00E1;mbito de especial riesgo por su incidencia en el desarrollo de las personas, en la configuraci&#x00F3;n del pensamiento y la acci&#x00F3;n a lo largo de la vida. Habr&#x00ED;a que enfatizar la precauci&#x00F3;n, que aparece ahora como exigencia de la no maleficencia, e incorporar el <italic>Principio Precautorio</italic> en aquellas intervenciones educativas donde la IA represente una amenaza para la autonom&#x00ED;a y la dignidad humana (ampliando la <xref ref-type="bibr" rid="ref-15-31821-ES">Declaraci&#x00F3;n de Wingspread</xref>).</p>
</sec>
</sec>
<sec id="sec-9-31821-ES" sec-type="conclusions|discussion">
<label><bold>7.</bold></label>
<title><bold>C<sc>onclusi&#x00F3;n y discusi&#x00F3;n</sc></bold></title>
<p>La perspectiva externa sobre IA y su uso en educaci&#x00F3;n, atiende a la repercusi&#x00F3;n social y la responsabilidad en las pr&#x00E1;cticas mediadas por estos sistemas (<xref ref-type="bibr" rid="ref-6-31821-ES">Bearman <italic>et al</italic>., 2023</xref>) y destacan la importancia de la &#x00E9;tica en el debate contempor&#x00E1;neo de la educaci&#x00F3;n (<xref ref-type="bibr" rid="ref-28-31821-ES">Holmes <italic>et al</italic>., 2022</xref>). Aqu&#x00ED; se ha buscado un marco unificado para este problema actual y con recorrido futuro, donde Ciencia, Tecnolog&#x00ED;a y Sociedad se han tenido en cuenta.</p>
<p>Este marco constituye un punto de partida &#x00FA;til, ya que (i) puede orientar una regulaci&#x00F3;n coherente y bien elaborada, (ii) permite ubicar las buenas pr&#x00E1;cticas, (iii) indica las competencias que necesitan desarrollar los docentes, y (iv) ofrece criterios a los estudiantes para interaccionar con la IA y aprovechar su potencial en los contextos formativos. Pero es preciso seguir avanzando. La investigaci&#x00F3;n en este &#x00E1;mbito est&#x00E1; experimentando un incremento notable y conviene, en futuros esfuerzos, poner estos consensos en relaci&#x00F3;n con los resultados del metaan&#x00E1;lisis de una muestra mucho mayor de documentos. A su vez:</p>
<list list-type="simple">
<list-item><label>1)</label> <p>Los principios representan los valores que se han de realizar a trav&#x00E9;s de los fines (p. ej., no ser&#x00ED;a leg&#x00ED;timo un fin que de ser alcanzado pueda da&#x00F1;ar). Pero adem&#x00E1;s de procurar fines &#x00E9;ticos para la IA, es preciso determinar los medios para desarrollarla de manera &#x00E9;tica, porque una cosa es hacer cosas &#x00E9;ticas y otra hacerlas &#x00E9;ticamente (<xref ref-type="bibr" rid="ref-28-31821-ES">Holmes <italic>et al</italic>., 2022</xref>, p. 504) y aunque lo que se busque sea algo beneficioso, no todo puede estar permitido para alcanzarlo (pi&#x00E9;nsese en el uso indebido de los datos).</p></list-item>
<list-item><label>2)</label> <p>Los valores a los que remiten los principios son valores &#x201C;confiables&#x201D; y constituyen algunas de nuestras certezas como garant&#x00ED;a de la dignidad humana y la convivencia democr&#x00E1;tica. Pero (a) pueden conceptualizarse de maneras distintas, como se observa si se atiende a los distintos criterios de <italic>justicia</italic> que aparecen en los diferentes c&#x00F3;digos y (b) son valores que surgieron dentro de la tradici&#x00F3;n occidental y, para garantizar un alcance global, es preciso incorporar la perspectiva de otras regiones y culturas. Esto nos sit&#x00FA;a de nuevo ante el viejo debate del multiculturalismo. Como se&#x00F1;ala Adela Cortina, &#x201C;ante retos universales no cabe sino la respuesta de una actitud &#x00E9;tica universal, que tiene como horizonte para la toma de decisiones el bien universal&#x201D; (<xref ref-type="bibr" rid="ref-13-31821-ES">Cortina, 1997</xref>, p. 261).</p></list-item>
<list-item><label>3)</label> <p>Pocas veces se concretan los mecanismos efectivos para que todas las partes interesadas asuman esos principios o la responsabilidad de su transgresi&#x00F3;n. Esto ata&#x00F1;e de modo especial a las empresas tecnol&#x00F3;gicas que desarrollan los productos de IA para la educaci&#x00F3;n y que adem&#x00E1;s controlan la investigaci&#x00F3;n actual (<xref ref-type="bibr" rid="ref-1-31821-ES">Ahmed <italic>et al</italic>., 2023</xref>). Como sujetos sociales, pueden y tienen que asumir la responsabilidad de sus acciones en la medida que: a) muestran intencionalidad en la actividad que despliegan y b) la toma de decisiones es controlada por la propia organizaci&#x00F3;n (<xref ref-type="bibr" rid="ref-25-31821-ES">Gonz&#x00E1;lez, 2020</xref>).</p></list-item>
</list>
<p>Las preocupaciones suscitadas por la IA entroncan con cuestiones tradicionales de la &#x00E9;tica. En general, no son novedosas. Cambia el grado de relevancia, puesto que la IA conforma en buena medida c&#x00F3;mo interaccionamos con el mundo, modula las relaciones sociales e, incluso, incide en la configuraci&#x00F3;n de nuestra propia identidad. Esto tiene consecuencias a la hora de considerar las nociones tradicionales de moralidad y responsabilidad individual, que quiz&#x00E1; necesiten ser revisadas a la luz de las nuevas condiciones.</p>
<p>El debate ha de plantearse en dos direcciones principales: Primera, la fundamentaci&#x00F3;n &#x00E9;tica que dote a los principios de legitimidad moral. Conviene evitar simplificaciones excesivas de cuestiones especialmente complejas, como sucede cuando son abordadas por expertos en IA. Decidir qu&#x00E9; es lo que no puede sacrificarse o qu&#x00E9; se debe preferir, requiere poder ofrecer razones s&#x00F3;lidas acerca de lo que se considera &#x201C;bueno&#x201D; o de lo que se &#x201C;debe hacer&#x201D;. Y esto es tarea de la &#x00E9;tica.</p>
<p>Para que el desarrollo y uso de los SIA en educaci&#x00F3;n se realice conforme a los principios, hace falta tambi&#x00E9;n una regulaci&#x00F3;n jur&#x00ED;dica y pol&#x00ED;tica. Pero las leyes han de conectar con los valores y estar en sinton&#x00ED;a con la raz&#x00F3;n y los deseos de las personas. Porque una cosa es una ley (que puede ser impuesta o asumida en el proceso de socializaci&#x00F3;n) y otra las razones que un ser humano tenga para tenerla por suya (<xref ref-type="bibr" rid="ref-13-31821-ES">Cortina, 1997</xref>). Entonces hay que considerar una educaci&#x00F3;n en valores.</p>
<p>Segunda, la determinaci&#x00F3;n del nuevo perfil profesional docente, que requiere precisar el conocimiento experto que ahora es necesario para el ejercicio de la profesi&#x00F3;n. Esto ha de incorporarse a los planes de formaci&#x00F3;n inicial del profesorado. Sobre la base de lo aqu&#x00ED; expuesto, deber&#x00E1; contemplar al menos: (a) formaci&#x00F3;n investigadora cient&#x00ED;ficamente s&#x00F3;lida que combine el rigor con la resoluci&#x00F3;n pr&#x00E1;ctica de problemas; (b) formaci&#x00F3;n &#x00E9;tica para abordar los desaf&#x00ED;os de la IA; y (c) formaci&#x00F3;n en IA que permita a los educadores ajustar su enfoque y dise&#x00F1;ar estrategias pedag&#x00F3;gicas m&#x00E1;s efectivas para avanzar hacia el objetivo de una &#x201C;educaci&#x00F3;n inclusiva, equitativa y de calidad y promover las oportunidades de aprendizaje durante toda la vida para todos&#x201D;.</p>
</sec>
</body>
<back>
<fn-group>
<fn id="fn1-ES" fn-type="other"><label><sup>1</sup></label> <p>Principios de Asilomar sobre IA (2017); Declaraci&#x00F3;n de Montreal para una IA Responsable (2017); Segunda versi&#x00F3;n de Dise&#x00F1;o &#x00C9;ticamente Alineado (2017); Declaraci&#x00F3;n sobre Inteligencia Artificial, Rob&#x00F3;tica y Sistemas &#x2018;Aut&#x00F3;nomos&#x2019;, del Grupo Europeo de &#x00C9;tica en la Ciencia y las Nuevas Tecnolog&#x00ED;as de la Comisi&#x00F3;n Europea (2018); Informe del Comit&#x00E9; de Inteligencia Artificial de la C&#x00E1;mara de los Lores del Reino Unido (2018); Asociaci&#x00F3;n sobre IA de m&#x00FA;ltiples partes interesadas (2018).</p></fn>
    <fn id="fn2-ES" fn-type="other"><label><sup>2</sup></label> <p>&#x201C;Iuris praecepta sunt haec: honeste vivere, alterum non laedere, suum cuique tribuere&#x201D; (<xref ref-type="bibr" rid="ref-43-31821-ES">Ulpiano</xref>. Digesto 1, 1, 10, 1).</p></fn>
</fn-group>
<ref-list>
<title><bold>R<sc>eferencias bibliogr&#x00E1;ficas</sc></bold></title>
<ref id="ref-1-31821-ES"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Ahmed</surname> <given-names>N.</given-names></name> <name><surname>Wahed</surname> <given-names>M.</given-names></name> <name><surname>Thompson</surname> <given-names>N. C.</given-names></name></person-group> <year>2023</year> <article-title>The growing influence of industry in AI research</article-title> <source>Science</source> <volume>379</volume> <issue>6635</issue> <fpage>884</fpage><lpage>886</lpage> <ext-link ext-link-type="uri" xlink:href="https://www.science.org/doi/10.1126/science.ade2420">https://www.science.org/doi/10.1126/science.ade2420</ext-link></element-citation>
<mixed-citation>Ahmed, N., Wahed, M., &#x0026; Thompson, N. C. (2023). The growing influence of industry in AI research. <italic>Science, 379</italic> (6635), 884-886. <ext-link ext-link-type="uri" xlink:href="https://www.science.org/doi/10.1126/science.ade2420">https://www.science.org/doi/10.1126/science.ade2420</ext-link></mixed-citation></ref>
<ref id="ref-2-31821-ES"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Aiken</surname> <given-names>R. M.</given-names></name> <name><surname>Epstein</surname> <given-names>R. G.</given-names></name></person-group> <year>2000</year> <article-title>Ethical Guidelines for AI in Education: Starting a Conversation</article-title> <source>International Journal of Artificial Intelligence in Education</source> <volume>11</volume> <fpage>163</fpage><lpage>176</lpage></element-citation>
<mixed-citation>Aiken, R. M., &#x0026; Epstein, R. G. (2000). Ethical Guidelines for AI in Education: Starting a Conversation. <italic>International Journal of Artificial Intelligence in Education, 11</italic>, 163-176.</mixed-citation></ref>
<ref id="ref-3-31821-ES"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Alonso</surname> <given-names>A. M.</given-names></name></person-group> <year>2022</year> <article-title>Las Ciencias de Dise&#x00F1;o como Ciencias Aplicadas y el papel de los valores: An&#x00E1;lisis del caso de la Educaci&#x00F3;n</article-title> <source>Revista de Investigaci&#x00F3;n Filos&#x00F3;fica</source> <volume>9</volume><issue>1</issue> <fpage>3</fpage><lpage>27</lpage> <pub-id pub-id-type="doi">10.26754/ojs_arif/arif.202216381</pub-id></element-citation>
<mixed-citation>Alonso, A. M. (2022). Las Ciencias de Dise&#x00F1;o como Ciencias Aplicadas y el papel de los valores: An&#x00E1;lisis del caso de la Educaci&#x00F3;n<italic>. Revista de Investigaci&#x00F3;n Filos&#x00F3;fica, 9</italic>(1), 3-27. <ext-link ext-link-type="uri" xlink:href="https://doi.org/10.26754/ojs_arif/arif.202216381">https://doi.org/10.26754/ojs_arif/arif.202216381</ext-link></mixed-citation></ref>
<ref id="ref-4-31821-ES"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Alonso</surname> <given-names>A. M.</given-names></name></person-group> <year>2023</year> <source>Filosof&#x00ED;a de la Ciencia de la Educaci&#x00F3;n: An&#x00E1;lisis como Ciencia Aplicada de Dise&#x00F1;o</source> <publisher-name>Tirant lo Blanch</publisher-name></element-citation>
<mixed-citation>Alonso, A. M. (2023). <italic>Filosof&#x00ED;a de la Ciencia de la Educaci&#x00F3;n: An&#x00E1;lisis como Ciencia Aplicada de Dise&#x00F1;o</italic>. Tirant lo Blanch.</mixed-citation></ref>
<ref id="ref-5-31821-ES"><element-citation publication-type="webpage"><person-group person-group-type="author"><collab>Asamblea General de la ONU</collab></person-group> <year>1948</year> <source>Declaraci&#x00F3;n Universal de los Derechos Humanos</source> <publisher-loc>Paris</publisher-loc> <ext-link ext-link-type="uri" xlink:href="http://www.un.org/en/universal-declaration-human-rights/">http://www.un.org/en/universal-declaration-human-rights/</ext-link></element-citation>
<mixed-citation>Asamblea General de la ONU. (1948). Declaraci&#x00F3;n Universal de los Derechos Humanos. Paris. <ext-link ext-link-type="uri" xlink:href="http://www.un.org/en/universal-declaration-human-rights/">http://www.un.org/en/universal-declaration-human-rights/</ext-link></mixed-citation></ref>
<ref id="ref-6-31821-ES"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Bearman</surname> <given-names>M.</given-names></name> <name><surname>Ryan</surname> <given-names>J.</given-names></name> <name><surname>Ajjawi</surname> <given-names>R.</given-names></name></person-group> <year>2023</year> <article-title>Discourses of artificial intelligence in higher education: a critical literature review</article-title> <source>Higher Education</source> <volume>86</volume> <fpage>369</fpage><lpage>385</lpage> <pub-id pub-id-type="doi">10.1007/s10734-022-00937-2</pub-id></element-citation>
<mixed-citation>Bearman, M., Ryan, J., &#x0026; Ajjawi, R. (2023). Discourses of artificial intelligence in higher education: a critical literature review. <italic>Higher Education, 86</italic>, 369&#x2013;385. <ext-link ext-link-type="uri" xlink:href="https://doi.org/10.1007/s10734-022-00937-2">https://doi.org/10.1007/s10734-022-00937-2</ext-link></mixed-citation></ref>
<ref id="ref-7-31821-ES"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Beauchamp</surname> <given-names>T. L.</given-names></name> <name><surname>Childress</surname> <given-names>J. F.</given-names></name></person-group> <year>2019</year> <source>Principles of biomedical ethics</source> <publisher-name>Oxford University Press</publisher-name></element-citation>
<mixed-citation>Beauchamp, T. L., &#x0026; Childress, J. F. (2019). <italic>Principles of biomedical ethics</italic>. Oxford University Press.</mixed-citation></ref>
<ref id="ref-8-31821-ES"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Boddington</surname> <given-names>P.</given-names></name></person-group> <year>2017</year> <source>Towards a code of ethics for artificial intelligence</source> <publisher-name>Springer</publisher-name></element-citation>
<mixed-citation>Boddington, P. (2017). <italic>Towards a code of ethics for artificial intelligence</italic>. Springer.</mixed-citation></ref>
<ref id="ref-9-31821-ES"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Boddington</surname> <given-names>P.</given-names></name></person-group> <year>2023</year> <chapter-title>Introduction, Why AI Ethics?</chapter-title> <comment>In</comment> <person-group person-group-type="author"><name><surname>Boddington</surname> <given-names>P.</given-names></name></person-group> <source>AI Ethics. A Textbook</source> <comment>pp.</comment> <fpage>1</fpage><lpage>34</lpage> <publisher-name>Springer</publisher-name></element-citation>
<mixed-citation>Boddington, P. (2023). Introduction, Why AI Ethics? In P. Boddington, <italic>AI Ethics. A Textbook</italic> (pp. 1-34). Springer.</mixed-citation></ref>
<ref id="ref-10-31821-ES"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Chu</surname> <given-names>J.</given-names></name> <name><surname>Xi</surname> <given-names>L.</given-names></name> <name><surname>Zhang</surname> <given-names>Q.</given-names></name> <name><surname>Lin</surname> <given-names>R.</given-names></name></person-group> <year>2022</year> <chapter-title>Research on Ethical Issues of Artificial Intelligence in Education</chapter-title> <comment>In</comment> <person-group person-group-type="editor"><name><surname>Yang</surname> <given-names>J.</given-names></name> <etal>et al.</etal></person-group> <role>Eds.</role> <source>Resilience and Future of Smart Learning</source> <comment>pp.</comment> <fpage>101</fpage><lpage>108</lpage> <publisher-name>Springer</publisher-name> <pub-id pub-id-type="doi">10.1007/978-981-19-5967-7_12</pub-id></element-citation>
<mixed-citation>Chu, J., Xi, L., Zhang, Q., &#x0026; Lin, R. (2022). Research on Ethical Issues of Artificial Intelligence in Education. In J. Yang, <italic>et al.</italic> (Eds.), <italic>Resilience and Future of Smart Learning</italic> (pp. 101-108). Springer. <ext-link ext-link-type="uri" xlink:href="https://doi.org/10.1007/978-981-19-5967-7_12">https://doi.org/10.1007/978-981-19-5967-7_12</ext-link></mixed-citation></ref>
<ref id="ref-11-31821-ES"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Coeckelberg</surname> <given-names>M.</given-names></name></person-group> <year>2021</year> <source>&#x00C9;tica de la Inteligencia Artificial</source> <publisher-name>C&#x00E1;tedra</publisher-name></element-citation>
<mixed-citation>Coeckelberg, M. (2021). <italic>&#x00C9;tica de la Inteligencia Artificial</italic>. C&#x00E1;tedra.</mixed-citation></ref>
<ref id="ref-12-31821-ES"><element-citation publication-type="webpage"><person-group person-group-type="author"><collab>Comisi&#x00F3;n Europea</collab></person-group> <year>2019</year> <article-title>Grupo de Expertos de Alto Nivel sobre IA</article-title> <source>Directrices &#x00C9;ticas para una IA fiable</source> <ext-link ext-link-type="uri" xlink:href="https://op.europa.eu/es/publication-detail/-/publication/d3988569-0434-11ea-8c1f-01aa75ed71a1">https://op.europa.eu/es/publication-detail/-/publication/d3988569-0434-11ea-8c1f-01aa75ed71a1</ext-link></element-citation>
<mixed-citation>Comisi&#x00F3;n Europea. (2019). Grupo de Expertos de Alto Nivel sobre IA. <italic>Directrices &#x00C9;ticas para una IA fiable</italic>. <ext-link ext-link-type="uri" xlink:href="https://op.europa.eu/es/publication-detail/-/publication/d3988569-0434-11ea-8c1f-01aa75ed71a1">https://op.europa.eu/es/publication-detail/-/publication/d3988569-0434-11ea-8c1f-01aa75ed71a1</ext-link></mixed-citation></ref>
<ref id="ref-13-31821-ES"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Cortina</surname> <given-names>A.</given-names></name></person-group> <year>1997</year> <source>Ciudadanos del mundo. Hacia una teor&#x00ED;a de la ciudadan&#x00ED;a</source> <publisher-name>Alianza editorial</publisher-name></element-citation>
<mixed-citation>Cortina, A. (1997). <italic>Ciudadanos del mundo. Hacia una teor&#x00ED;a de la ciudadan&#x00ED;a.</italic> Alianza editorial.</mixed-citation></ref>
<ref id="ref-14-31821-ES"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Crawford</surname> <given-names>K.</given-names></name></person-group> <year>2023</year> <source>Atlas de I. Poder, Pol&#x00ED;tica y Costes Planetarios de la Inteligencia Artificial</source> <publisher-name>Ned ediciones</publisher-name></element-citation>
<mixed-citation>Crawford, K. (2023). <italic>Atlas de I. Poder, Pol&#x00ED;tica y Costes Planetarios de la Inteligencia Artificial</italic>. Ned ediciones.</mixed-citation></ref>
<ref id="ref-15-31821-ES"><element-citation publication-type="webpage"><person-group person-group-type="author"><collab>Declaraci&#x00F3;n de Wingspread sobre el Principio Precautorio</collab></person-group> <year>1998</year> <ext-link ext-link-type="uri" xlink:href="http://www.linea-e.com/cuadernos/pdfs/numero11/elprincipiodeprecaucion.pdf">http://www.linea-e.com/cuadernos/pdfs/numero11/elprincipiodeprecaucion.pdf</ext-link></element-citation>
<mixed-citation>Declaraci&#x00F3;n de Wingspread sobre el Principio Precautorio (1998). <ext-link ext-link-type="uri" xlink:href="http://www.linea-e.com/cuadernos/pdfs/numero11/elprincipiodeprecaucion.pdf">http://www.linea-e.com/cuadernos/pdfs/numero11/elprincipiodeprecaucion.pdf</ext-link></mixed-citation></ref>
<ref id="ref-16-31821-ES"><element-citation publication-type="webpage"><person-group person-group-type="author"><collab>Executive Office of the President</collab></person-group> <year>2020</year> <source>Guidance for regulation of Artificial Intelligence applications</source> <ext-link ext-link-type="uri" xlink:href="https://www.whitehouse.gov/wp-content/uploads/2020/11/M-21-06.pdf">https://www.whitehouse.gov/wp-content/uploads/2020/11/M-21-06.pdf</ext-link></element-citation>
<mixed-citation>Executive Office of the President. (2020). <italic>Guidance for regulation of Artificial Intelligence applications</italic>. <ext-link ext-link-type="uri" xlink:href="https://www.whitehouse.gov/wp-content/uploads/2020/11/M-21-06.pdf">https://www.whitehouse.gov/wp-content/uploads/2020/11/M-21-06.pdf</ext-link></mixed-citation></ref>
<ref id="ref-17-31821-ES"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Flores Vivar</surname> <given-names>J. M.</given-names></name> <name><surname>Garc&#x00ED;a-Pe&#x00F1;alvo</surname> <given-names>F. J.</given-names></name></person-group> <year>2023</year> <article-title>Reflexiones sobre la &#x00E9;tica, potencialidades y retos de la Inteligencia Artificial en el marco de la Educaci&#x00F3;n de calidad (ODS4)</article-title> <source>Comunicar</source> <volume>31</volume><issue>74</issue> <fpage>37</fpage><lpage>47</lpage> <pub-id pub-id-type="doi">10.3916/C74-2023-03</pub-id></element-citation>
<mixed-citation>Flores Vivar, J. M. y Garc&#x00ED;a-Pe&#x00F1;alvo, F. J. (2023). Reflexiones sobre la &#x00E9;tica, potencialidades y retos de la Inteligencia Artificial en el marco de la Educaci&#x00F3;n de calidad (ODS4). <italic>Comunicar, 31</italic>(74), 37-47. <ext-link ext-link-type="uri" xlink:href="https://doi.org/10.3916/C74-2023-03">https://doi.org/10.3916/C74-2023-03</ext-link></mixed-citation></ref>
<ref id="ref-18-31821-ES"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Floridi</surname> <given-names>L.</given-names></name></person-group> <year>2014</year> <source>The Fourth Revolution - How the Infosphere is Reshaping Human Reality</source> <publisher-name>Oxford University Press</publisher-name></element-citation>
<mixed-citation>Floridi, L. (2014). <italic>The Fourth Revolution - How the Infosphere is Reshaping Human Reality.</italic> Oxford University Press.</mixed-citation></ref>
<ref id="ref-19-31821-ES"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Floridi</surname> <given-names>L</given-names></name></person-group> <year>2022</year> <source>Etica dell&#x2019;intelligenza artificiale. Sviluppi, opportunit&#x00E0;, sfide</source> <publisher-name>Rafaello Cortina</publisher-name></element-citation>
<mixed-citation>Floridi, L (2022). <italic>Etica dell&#x2019;intelligenza artificiale. Sviluppi, opportunit&#x00E0;, sfide.</italic> Rafaello Cortina.</mixed-citation></ref>
<ref id="ref-20-31821-ES"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Floridi</surname> <given-names>L.</given-names></name> <name><surname>Cowls</surname> <given-names>J.</given-names></name></person-group> <year>2021</year> <chapter-title>A Unified Framework of Five Principles for AI in Society</chapter-title> <comment>In</comment> <person-group person-group-type="editor"><name><surname>Floridi</surname> <given-names>L.</given-names></name></person-group> <role>Ed.</role> <source>Ethics, Governance, and Policies in Artificial Intelligence</source> <fpage>5</fpage><lpage>17</lpage> <publisher-name>Springer</publisher-name></element-citation>
<mixed-citation>Floridi L., &#x0026; Cowls, J. (2021). A Unified Framework of Five Principles for AI in Society. In L. Floridi (Ed.), <italic>Ethics, Governance, and Policies in Artificial Intelligence</italic> (5-17). Springer.</mixed-citation></ref>
<ref id="ref-21-31821-ES"><element-citation publication-type="webpage"><person-group person-group-type="author"><collab>Gobierno de Espa&#x00F1;a</collab></person-group> <year>2020</year> <source>Estrategia Nacional de Inteligencia Artificial</source> <ext-link ext-link-type="uri" xlink:href="https://www.lamoncloa.gob.es/presidente/actividades/Documents/2020/021220-ENIA.pdf">https://www.lamoncloa.gob.es/presidente/actividades/Documents/2020/021220-ENIA.pdf</ext-link></element-citation>
<mixed-citation>Gobierno de Espa&#x00F1;a. (2020). <italic>Estrategia Nacional de Inteligencia Artificial</italic>. <ext-link ext-link-type="uri" xlink:href="https://www.lamoncloa.gob.es/presidente/actividades/Documents/2020/021220-ENIA.pdf">https://www.lamoncloa.gob.es/presidente/actividades/Documents/2020/021220-ENIA.pdf</ext-link></mixed-citation></ref>
<ref id="ref-22-31821-ES"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Gonz&#x00E1;lez</surname> <given-names>W. J.</given-names></name></person-group> <year>2007</year> <chapter-title>Configuraci&#x00F3;n de las Ciencias de Dise&#x00F1;o como Ciencias de lo Artificial: Papel de la Inteligencia Artificial y de la racionalidad limitada</chapter-title> <comment>En</comment> <person-group person-group-type="editor"><name><surname>Gonz&#x00E1;lez</surname> <given-names>W. J.</given-names></name></person-group> <role>Ed.</role> <source>Las Ciencias de Dise&#x00F1;o: Racionalidad limitada, predicci&#x00F3;n y prescripci&#x00F3;n</source> <fpage>41</fpage><lpage>69</lpage> <publisher-name>Netbiblo</publisher-name></element-citation>
<mixed-citation>Gonz&#x00E1;lez, W. J. (2007). Configuraci&#x00F3;n de las Ciencias de Dise&#x00F1;o como Ciencias de lo Artificial: Papel de la Inteligencia Artificial y de la racionalidad limitada. En W. J. Gonz&#x00E1;lez (Ed.), <italic>Las Ciencias de Dise&#x00F1;o: Racionalidad limitada, predicci&#x00F3;n y prescripci&#x00F3;n</italic> (41-69). Netbiblo.</mixed-citation></ref>
<ref id="ref-23-31821-ES"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Gonz&#x00E1;lez</surname> <given-names>W. J.</given-names></name></person-group> <year>2017a</year> <article-title>Artificial Intelligence in a New Context: &#x201C;Internal&#x201D; and &#x201C;External&#x201D; Factors</article-title> <source>Minds &#x0026; Machines</source> <volume>27</volume><issue>3</issue> <fpage>393</fpage><lpage>396</lpage> <pub-id pub-id-type="doi">10.1007/s11023-017-9444-3</pub-id></element-citation>
<mixed-citation>Gonz&#x00E1;lez, W. J. (2017a). Artificial Intelligence in a New Context: &#x201C;Internal&#x201D; and &#x201C;External&#x201D; Factors. <italic>Minds &#x0026; Machines, 27</italic>(3), 393&#x2013;396. <ext-link ext-link-type="uri" xlink:href="https://doi.org/10.1007/s11023-017-9444-3">https://doi.org/10.1007/s11023-017-9444-3</ext-link></mixed-citation></ref>
<ref id="ref-24-31821-ES"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Gonz&#x00E1;lez</surname> <given-names>W. J.</given-names></name></person-group> <year>2017b</year> <article-title>From Intelligence to Rationality of Minds and Machines in Contemporary Society: The Sciences of Design and the Role of Information</article-title> <source>Minds &#x0026; Machines</source> <volume>27</volume><issue>3</issue> <fpage>397</fpage><lpage>424</lpage> <pub-id pub-id-type="doi">10.1007/s11023-017-9439-0</pub-id></element-citation>
<mixed-citation>Gonz&#x00E1;lez, W. J. (2017b). From Intelligence to Rationality of Minds and Machines in Contemporary Society: The Sciences of Design and the Role of Information. <italic>Minds &#x0026; Machines, 27</italic>(3), 397-424. <ext-link ext-link-type="uri" xlink:href="https://doi.org/10.1007/s11023-017-9439-0">https://doi.org/10.1007/s11023-017-9439-0</ext-link></mixed-citation></ref>
<ref id="ref-25-31821-ES"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Gonz&#x00E1;lez</surname> <given-names>W. J.</given-names></name></person-group> <year>2020</year> <article-title>The Internet at the Service of Society: Business Ethics, Rationality, and Responsibility</article-title> <source>&#x00C9;ndoxa</source> <volume>46</volume> <fpage>383</fpage><lpage>412</lpage> <pub-id pub-id-type="doi">10.5944/endoxa.46.2020.28029</pub-id></element-citation>
<mixed-citation>Gonz&#x00E1;lez, W. J. (2020). The Internet at the Service of Society: Business Ethics, Rationality, and Responsibility. <italic>&#x00C9;ndoxa, 46</italic>, 383-412. <ext-link ext-link-type="uri" xlink:href="https://doi.org/10.5944/endoxa.46.2020.28029">https://doi.org/10.5944/endoxa.46.2020.28029</ext-link></mixed-citation></ref>
<ref id="ref-26-31821-ES"><element-citation publication-type="webpage"><person-group person-group-type="author"><collab>Google LLC</collab></person-group> <year>2022</year> <source>Artificial Intelligence at Google: Our Principles</source> <ext-link ext-link-type="uri" xlink:href="https://ai.google/principles/">https://ai.google/principles/</ext-link></element-citation>
<mixed-citation>Google LLC. (2022). <italic>Artificial Intelligence at Google: Our Principles</italic>. <ext-link ext-link-type="uri" xlink:href="https://ai.google/principles/">https://ai.google/principles/</ext-link></mixed-citation></ref>
<ref id="ref-27-31821-ES"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Holmes</surname> <given-names>W.</given-names></name> <name><surname>Porayska-Pomsta</surname> <given-names>K.</given-names></name></person-group> <year>2023</year> <source>The Ethics of Artificial Intelligence in Education</source> <publisher-name>Routledge Taylor</publisher-name></element-citation>
<mixed-citation>Holmes, W., &#x0026; Porayska-Pomsta, K. (2023). <italic>The Ethics of Artificial Intelligence in Education.</italic> Routledge Taylor.</mixed-citation></ref>
<ref id="ref-28-31821-ES"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Holmes</surname> <given-names>W.</given-names></name> <name><surname>Porayska-Pomsta</surname> <given-names>K.</given-names></name> <name><surname>Holstein</surname> <given-names>K.</given-names></name> <etal>et al.</etal></person-group> <year>2022</year> <article-title>Ethics of AI in Education: Towards a Community-Wide Framework</article-title> <source>International Journal of Artificial Intelligence in Education</source> <volume>32</volume> <fpage>504</fpage><lpage>526</lpage> <pub-id pub-id-type="doi">10.1007/s40593-021-00239-1</pub-id></element-citation>
<mixed-citation>Holmes, W., Porayska-Pomsta, K., Holstein, K. et al. (2022). Ethics of AI in Education: Towards a Community-Wide Framework. <italic>International Journal of Artificial Intelligence in Education, 32</italic>, 504&#x2013;526. <ext-link ext-link-type="uri" xlink:href="https://doi.org/10.1007/s40593-021-00239-1">https://doi.org/10.1007/s40593-021-00239-1</ext-link></mixed-citation></ref>
<ref id="ref-29-31821-ES"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Humble</surname> <given-names>N.</given-names></name> <name><surname>Mozellius</surname> <given-names>P.</given-names></name></person-group> <year>2022</year> <article-title>The threat, hype, and promise of artificial intelligence in education</article-title> <source>Discover Artificial Intelligence</source> <volume>2</volume><issue>22</issue> <pub-id pub-id-type="doi">10.1007/s44163-022-00039-z</pub-id></element-citation>
<mixed-citation>Humble, N., &#x0026; Mozellius, P. (2022). The threat, hype, and promise of artificial intelligence in education. <italic>Discover Artificial Intelligence, 2</italic>(22). <ext-link ext-link-type="uri" xlink:href="https://doi.org/10.1007/s44163-022-00039-z">https://doi.org/10.1007/s44163-022-00039-z</ext-link></mixed-citation></ref>
<ref id="ref-30-31821-ES"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Jara</surname> <given-names>I.</given-names></name> <name><surname>Ochoa</surname> <given-names>J. M.</given-names></name></person-group> <year>2020</year> <source>Usos y efectos de la inteligencia artificial en educaci&#x00F3;n</source> <publisher-name>Banco Interamericano de Desarrollo</publisher-name> <pub-id pub-id-type="doi">10.18235/0002380</pub-id></element-citation>
<mixed-citation>Jara, I., &#x0026; Ochoa, J. M. (2020). <italic>Usos y efectos de la inteligencia artificial en educaci&#x00F3;n</italic>. Banco Interamericano de Desarrollo. <ext-link ext-link-type="uri" xlink:href="http://dx.doi.org/10.18235/0002380">http://dx.doi.org/10.18235/0002380</ext-link></mixed-citation></ref>
<ref id="ref-31-31821-ES"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Linares Salgado</surname> <given-names>J. E.</given-names></name></person-group> <year>2022</year> <article-title>Principios &#x00E9;ticos para el desarrollo de la Inteligencia Artificial y su aplicaci&#x00F3;n en los sistemas de salud</article-title> <source>ArtefaCToS. Revista de Estudios de la Ciencia y la Tecnolog&#x00ED;a</source> <volume>11</volume><issue>2</issue> <fpage>137</fpage><lpage>161</lpage></element-citation>
<mixed-citation>Linares Salgado, J. E. (2022). Principios &#x00E9;ticos para el desarrollo de la Inteligencia Artificial y su aplicaci&#x00F3;n en los sistemas de salud. <italic>ArtefaCToS. Revista de Estudios de la Ciencia y la Tecnolog&#x00ED;a, 11</italic>(2), 137-161.</mixed-citation></ref>
<ref id="ref-32-31821-ES"><element-citation publication-type="webpage"><person-group person-group-type="author"><collab>Meta Platforms</collab></person-group> <year>2021</year> <source>Facebook five pillars of responsible AI</source> <ext-link ext-link-type="uri" xlink:href="https://ai.facebook.com/blog/facebooks-five-pillars-of-responsible-ai/">https://ai.facebook.com/blog/facebooks-five-pillars-of-responsible-ai/</ext-link></element-citation>
<mixed-citation>Meta Platforms. (2021). <italic>Facebook five pillars of responsible AI</italic>. <ext-link ext-link-type="uri" xlink:href="https://ai.facebook.com/blog/facebooks-five-pillars-of-responsible-ai/">https://ai.facebook.com/blog/facebooks-five-pillars-of-responsible-ai/</ext-link></mixed-citation></ref>
<ref id="ref-33-31821-ES"><element-citation publication-type="webpage"><person-group person-group-type="author"><collab>Microsoft Corporation</collab></person-group> <year>2022</year> <source>Responsible IA</source> <ext-link ext-link-type="uri" xlink:href="https://www.microsoft.com/en-us/ai/responsible-ai?activetab=pivot1%3aprimaryr6">https://www.microsoft.com/en-us/ai/responsible-ai?activetab=pivot1%3aprimaryr6</ext-link></element-citation>
<mixed-citation>Microsoft Corporation (2022). <italic>Responsible IA</italic>. <ext-link ext-link-type="uri" xlink:href="https://www.microsoft.com/en-us/ai/responsible-ai?activetab=pivot1%3aprimaryr6">https://www.microsoft.com/en-us/ai/responsible-ai?activetab=pivot1%3aprimaryr6</ext-link></mixed-citation></ref>
<ref id="ref-34-31821-ES"><element-citation publication-type="webpage"><person-group person-group-type="author"><collab>Ministerio de Ciencia y Tecnolog&#x00ED;a de la Rep&#x00FA;blica Popular China</collab></person-group> <year>2017</year> <source>Principios de Gobernanza para una Nueva Generaci&#x00F3;n de IA: Desarrollo Responsable de la Inteligencia Artificial</source> <ext-link ext-link-type="uri" xlink:href="https://www.newamerica.org/cybersecurity-initiative/digichina/blog/translation-chinese-expert-group-offers-governance-principles-responsible-ai/">https://www.newamerica.org/cybersecurity-initiative/digichina/blog/translation-chinese-expert-group-offers-governance-principles-responsible-ai/</ext-link></element-citation>
<mixed-citation>Ministerio de Ciencia y Tecnolog&#x00ED;a de la Rep&#x00FA;blica Popular China. (2017). <italic>Principios de Gobernanza para una Nueva Generaci&#x00F3;n de IA: Desarrollo Responsable de la Inteligencia Artificial</italic>. <ext-link ext-link-type="uri" xlink:href="https://www.newamerica.org/cybersecurity-initiative/digichina/blog/translation-chinese-expert-group-offers-governance-principles-responsible-ai/">https://www.newamerica.org/cybersecurity-initiative/digichina/blog/translation-chinese-expert-group-offers-governance-principles-responsible-ai/</ext-link></mixed-citation></ref>
<ref id="ref-35-31821-ES"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Moreno Padilla</surname> <given-names>R.</given-names></name></person-group> <year>2019</year> <article-title>La llegada de la inteligencia artificial a la educaci&#x00F3;n</article-title> <source>RITI Journal</source> <volume>7</volume><issue>14</issue> <fpage>260</fpage><lpage>270</lpage> <pub-id pub-id-type="doi">10.36825/RITI.07.14.022</pub-id></element-citation>
<mixed-citation>Moreno Padilla, R. (2019). La llegada de la inteligencia artificial a la educaci&#x00F3;n. <italic>RITI Journal, 7</italic>(14), 260-270. <ext-link ext-link-type="uri" xlink:href="https://doi.org/10.36825/RITI.07.14.022">https://doi.org/10.36825/RITI.07.14.022</ext-link></mixed-citation></ref>
<ref id="ref-36-31821-ES"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Nguyen</surname> <given-names>A</given-names></name> <name><surname>Ngo</surname> <given-names>H. N</given-names></name> <name><surname>Hong</surname> <given-names>Y.</given-names></name> <name><surname>Dang</surname> <given-names>B.</given-names></name> <name><surname>Nguyen</surname> <given-names>B.T.</given-names></name></person-group> <year>2023</year> <article-title>Ethical principles for artificial intelligence in education</article-title> <source>Educ Inf Technol</source> <volume>28</volume><issue>4</issue> <fpage>4221</fpage><lpage>4241</lpage> <pub-id pub-id-type="doi">10.1007/s10639-022-11316-w</pub-id></element-citation>
<mixed-citation>Nguyen A, Ngo H. N, Hong, Y., Dang, B., &#x0026; Nguyen, B.T. (2023). Ethical principles for artificial intelligence in education. <italic>Educ Inf Technol, 28</italic>(4), 4221-4241. <ext-link ext-link-type="uri" xlink:href="https://doi.org/10.1007/s10639-022-11316-w">https://doi.org/10.1007/s10639-022-11316-w</ext-link></mixed-citation></ref>
<ref id="ref-37-31821-ES"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Nussbaum</surname> <given-names>M.</given-names></name></person-group> <year>2012</year> <source>Crear capacidades. Propuesta para el desarrollo humano</source> <publisher-name>Paid&#x00F3;s</publisher-name></element-citation>
<mixed-citation>Nussbaum, M. (2012). <italic>Crear capacidades. Propuesta para el desarrollo humano.</italic> Paid&#x00F3;s.</mixed-citation></ref>
<ref id="ref-38-31821-ES"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Oca&#x00F1;a-Fernandez</surname> <given-names>Y.</given-names></name> <name><surname>Valenzuela-Fernandez</surname> <given-names>L.</given-names></name> <name><surname>Garro-Aburto</surname> <given-names>L.</given-names></name></person-group> <year>2019</year> <article-title>Inteligencia artificial y sus implicaciones en la educaci&#x00F3;n superior</article-title> <source>Prop&#x00F3;sitos y Representaciones</source> <volume>7</volume><issue>2</issue> <fpage>536</fpage><lpage>552</lpage> <pub-id pub-id-type="doi">10.20511/pyr2019.v7n2.274</pub-id></element-citation>
<mixed-citation>Oca&#x00F1;a-Fernandez, Y., Valenzuela-Fernandez, L., &#x0026; Garro-Aburto, L. (2019). Inteligencia artificial y sus implicaciones en la educaci&#x00F3;n superior. <italic>Prop&#x00F3;sitos y Representaciones, 7</italic>(2), 536-552. <ext-link ext-link-type="uri" xlink:href="http://dx.doi.org/10.20511/pyr2019.v7n2.274">http://dx.doi.org/10.20511/pyr2019.v7n2.274</ext-link></mixed-citation></ref>
<ref id="ref-39-31821-ES"><element-citation publication-type="webpage"><person-group person-group-type="author"><collab>OECD</collab></person-group> <year>2019</year> <source>Recommendation of the Council on Artificial Intelligence</source> <comment>OECD/LEGAL/0449</comment> <ext-link ext-link-type="uri" xlink:href="https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0449">https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0449</ext-link></element-citation>
<mixed-citation>OECD (2019). <italic>Recommendation of the Council on Artificial Intelligence</italic>. OECD/LEGAL/0449. <ext-link ext-link-type="uri" xlink:href="https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0449">https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0449</ext-link></mixed-citation></ref>
<ref id="ref-40-31821-ES"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>S&#x00E6;tra</surname> <given-names>H.S.</given-names></name> <name><surname>Danaher</surname> <given-names>J.</given-names></name></person-group> <year>2022</year> <article-title>To Each Technology Its Own Ethics: The Problem of Ethical Proliferation</article-title> <source>Philosophy &#x0026; Technology</source> <volume>35</volume> <fpage>93</fpage> <pub-id pub-id-type="doi">10.1007/s13347-022-00591-7</pub-id></element-citation>
<mixed-citation>S&#x00E6;tra, H.S., &#x0026; Danaher, J. (2022). To Each Technology Its Own Ethics: The Problem of Ethical Proliferation. <italic>Philosophy &#x0026; Technology, 35</italic>, 93. <ext-link ext-link-type="uri" xlink:href="https://doi.org/10.1007/s13347-022-00591-7">https://doi.org/10.1007/s13347-022-00591-7</ext-link></mixed-citation></ref>
<ref id="ref-41-31821-ES"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>S&#x00E1;nchez Vila</surname> <given-names>E. M.</given-names></name> <name><surname>Lama Pen&#x00ED;n</surname> <given-names>M.</given-names></name></person-group> <year>2007</year> <article-title>Monograf&#x00ED;a: T&#x00E9;cnicas de la Inteligencia Artificial Aplicadas a la Educaci&#x00F3;n</article-title> <source>Inteligencia Artificial. Revista Iberoamericana de Inteligencia Artificial</source> <volume>11</volume><issue>33</issue> <fpage>7</fpage><lpage>12</lpage></element-citation>
<mixed-citation>S&#x00E1;nchez Vila, E. M., &#x0026; Lama Pen&#x00ED;n, M. (2007). Monograf&#x00ED;a: T&#x00E9;cnicas de la Inteligencia Artificial Aplicadas a la Educaci&#x00F3;n. <italic>Inteligencia Artificial. Revista Iberoamericana de Inteligencia Artificial, 11</italic>(33), 7-12.</mixed-citation></ref>
<ref id="ref-42-31821-ES"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Simon</surname> <given-names>H. A.</given-names></name></person-group> <year>1996</year> <source>The Sciences of the Artificial</source> <publisher-name>The MIT Press</publisher-name> <edition>3&#x00AA; ed.</edition></element-citation>
<mixed-citation>Simon, H. A. (1996). <italic>The Sciences of the Artificial</italic>. The MIT Press. 3&#x00AA; ed.</mixed-citation></ref>
<ref id="ref-43-31821-ES"><element-citation publication-type="webpage"><person-group person-group-type="author"><name><surname>Ulpiano</surname> <given-names>E. D.</given-names></name></person-group> <year>1836</year> <article-title>Digesto</article-title> <comment>En</comment> <person-group person-group-type="author"><collab>Academicun parisiense</collab></person-group> <source>Corpus Iuris civilis</source> <ext-link ext-link-type="uri" xlink:href="https://gallica.bnf.fr/ark:/12148/bpt6k65768518/f13.item.r=digestum">https://gallica.bnf.fr/ark:/12148/bpt6k65768518/f13.item.r=digestum</ext-link></element-citation>
<mixed-citation>Ulpiano, E. D. (1836). Digesto. En Academicun parisiense, <italic>Corpus Iuris civilis</italic>. <ext-link ext-link-type="uri" xlink:href="https://gallica.bnf.fr/ark:/12148/bpt6k65768518/f13.item.r=digestum">https://gallica.bnf.fr/ark:/12148/bpt6k65768518/f13.item.r=digestum</ext-link></mixed-citation></ref>
<ref id="ref-44-31821-ES"><element-citation publication-type="webpage"><person-group person-group-type="author"><collab>Unesco</collab></person-group> <year>2019</year> <source>Beijing consensus on artificial intelligence and education</source> <ext-link ext-link-type="uri" xlink:href="https://unesdoc.unesco.org/ask:/48223/pf0000368303">https://unesdoc.unesco.org/ask:/48223/pf0000368303</ext-link></element-citation>
<mixed-citation>Unesco. (2019). <italic>Beijing consensus on artificial intelligence and education</italic>. <ext-link ext-link-type="uri" xlink:href="https://unesdoc.unesco.org/ask:/48223/pf0000368303">https://unesdoc.unesco.org/ask:/48223/pf0000368303</ext-link></mixed-citation></ref>
<ref id="ref-45-31821-ES"><element-citation publication-type="webpage"><person-group person-group-type="author"><collab>Unesco</collab></person-group> <year>2022</year> <source>Recomendaci&#x00F3;n sobre la &#x00E9;tica de la inteligencia artificial</source> <ext-link ext-link-type="uri" xlink:href="https://unesdoc.unesco.org/ark:/48223/pf0000379920_spa.page=15">https://unesdoc.unesco.org/ark:/48223/pf0000379920_spa.page=15</ext-link></element-citation>
<mixed-citation>Unesco. (2022). <italic>Recomendaci&#x00F3;n sobre la &#x00E9;tica de la inteligencia artificial</italic>. <ext-link ext-link-type="uri" xlink:href="https://unesdoc.unesco.org/ark:/48223/pf0000379920_spa.page=15">https://unesdoc.unesco.org/ark:/48223/pf0000379920_spa.page=15</ext-link>.</mixed-citation></ref>
<ref id="ref-46-31821-ES"><element-citation publication-type="webpage"><person-group person-group-type="author"><collab>Uni&#x00F3;n Europea</collab></person-group> <year>2022</year> <source>Directrices &#x00E9;ticas sobre el uso de la inteligencia artificial (IA) y los datos en la educaci&#x00F3;n y formaci&#x00F3;n para los educadores</source> <publisher-name>Publications Office of the European Union</publisher-name> <ext-link ext-link-type="uri" xlink:href="https://data.europa.eu/doi/10.2766/898">https://data.europa.eu/doi/10.2766/898</ext-link></element-citation>
<mixed-citation>Uni&#x00F3;n Europea. (2022). <italic>Directrices &#x00E9;ticas sobre el uso de la inteligencia artificial (IA) y los datos en la educaci&#x00F3;n y formaci&#x00F3;n para los educadores</italic>. Publications Office of the European Union. <ext-link ext-link-type="uri" xlink:href="https://data.europa.eu/doi/10.2766/898">https://data.europa.eu/doi/10.2766/898</ext-link></mixed-citation></ref>
<ref id="ref-47-31821-ES"><element-citation publication-type="report"><person-group person-group-type="author"><collab>United Nations</collab></person-group> <year>2019</year> <source>The Sustainable Development Goals Report</source> <volume>2019</volume> <ext-link ext-link-type="uri" xlink:href="https://bit.ly/34nbq60">https://bit.ly/34nbq60</ext-link></element-citation>
<mixed-citation>United Nations. (2019). <italic>The Sustainable Development Goals Report 2019</italic>. <ext-link ext-link-type="uri" xlink:href="https://bit.ly/34nbq60">https://bit.ly/34nbq60</ext-link></mixed-citation></ref>
</ref-list>
</back>
<sub-article article-type="translation" xml:lang="en" id="S1">
<front-stub>
<article-id pub-id-type="publisher-id">teri.31821</article-id>
<article-id pub-id-type="doi">10.14201/teri.31821</article-id>
<article-categories>
<subj-group subj-group-type="heading">
<subject>Monographic</subject>
</subj-group>
</article-categories>
<title-group>
    <article-title>Towards an Ethical Framework for Artificial Intelligence in Education</article-title>
<trans-title-group>
<trans-title xml:lang="es"><italic>Hacia un marco &#x00E9;tico de la inteligencia artificial en la educaci&#x00F3;n</italic></trans-title>
</trans-title-group>
</title-group>
<contrib-group>
<contrib contrib-type="author" corresp="yes">
<contrib-id contrib-id-type="orcid">https://orcid.org/0000-0002-2379-657X</contrib-id>
<name>
<surname>ALONSO-RODR&#x00CD;GUEZ</surname>
<given-names>Ana Mar&#x00ED;a</given-names>
</name>
<xref ref-type="aff" rid="aff1-EN"/>
<xref ref-type="corresp" rid="c1-EN"/>
</contrib>
<aff id="aff1-EN">
<institution content-type="original">Centre for Research in the Philosophy of Science and Technology (CIFYT). University of A Coru&#x00F1;a. Spain.</institution>
<institution content-type="orgname">Centre for Research in the Philosophy of Science and Technology (CIFYT)</institution>
<institution content-type="orgdiv1">University of A Coru&#x00F1;a</institution>
<country country="ES">Spain</country>
</aff>
</contrib-group>
<author-notes>
<corresp id="c1-EN"><email>a.alonsor@udc.es</email></corresp>
</author-notes>
<pub-date publication-format="electronic" date-type="pub">
    <day>04</day>
    <month>06</month>
<year>2024</year>
</pub-date>
<pub-date publication-format="electronic" date-type="collection">
    <day>01</day>
    <month>07</month>
<year>2024</year>
</pub-date>
<volume>36</volume>
<issue>2</issue>
<issue-id pub-id-type="doi">10.14201/teri.2024362</issue-id>
<fpage>79</fpage>
<lpage>98</lpage>
<history>
<date date-type="received">
<day>29</day>
<month>12</month>
<year>2023</year>
</date>
<date date-type="accepted">
<day>22</day>
<month>02</month>
<year>2024</year>
</date>
</history>
<permissions>
<copyright-statement>&#x00A9; 2024 Ediciones Universidad de Salamanca</copyright-statement>
<copyright-year>2024</copyright-year>
<license license-type="open-access" xlink:href="https://creativecommons.org/licenses/by-nc-nd/4.0/" xml:lang="en">

<license-p>Esta obra est&#x00E1; bajo una licencia internacional Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International</license-p>
</license>
</permissions>
<abstract>
<title>ABSTRACT</title>
<p>This article reflects on the use of Artificial Intelligence in education from an ethical perspective. It does so from an external perspective, considering its impact on educational contexts as a breeding ground for the ethical and political challenges that society faces. This allows us to measure their scope and depth and propose actions to address them.</p>
<p>The <italic>objectives</italic> of this article focus on the ethical bases of Artificial Intelligence related to educational activity, seeking to identify: a) the opportunities, associated risks and ethical impact on education; and b) the ethical principles that should guide the development, deployment and use of these intelligent systems.</p>
<p>To achieve these objectives, a qualitative study was carried out, supported by a literature review based on the following method: (i) academic works on the current uses and potential risks of Artificial Intelligence and (ii) a comparative analysis of different ethical codes, exploring the convergence of principles applicable to Artificial Intelligence Systems in educational contexts.</p>
<p>The results obtained, in the first instance, place the identified problems in the ethical tradition and question the proliferation of subdomains within the discipline. The possibility of a unified ethical framework that avoids the overlap of principles for each specific domain is then investigated. The findings confirm the usefulness of a widely recognized and influential framework with principles that adapt well to the ethical challenges of education.</p>
<p>It concludes by indicating potential lines of future research: (I) ethical foundation and normative regulation for the development and use of Artificial Intelligence in education in accordance with the selected principles; and (II) definition of a new professional teaching profile and its implications for initial teacher training.</p>
</abstract>
<trans-abstract xml:lang="es">
<title>RESUMEN</title>
<p>Este art&#x00ED;culo reflexiona sobre el uso de la Inteligencia Artificial en educaci&#x00F3;n desde una perspectiva &#x00E9;tica. Lo hace desde un punto de vista externo, considerando su incidencia en los contextos educativos como vivero de los desaf&#x00ED;os &#x00E9;ticos y pol&#x00ED;ticos que encara la sociedad. Esto permite dimensionar su alcance y profundidad y proponer medidas para afrontarlos.</p>
<p>Sus <italic>objetivos</italic> se enfocan hacia las bases &#x00E9;ticas de la Inteligencia Artificial relacionada con la actividad educativa, buscando identificar: las oportunidades, riesgos asociados y su impacto &#x00E9;tico en educaci&#x00F3;n; y b) los principios &#x00E9;ticos que puedan guiar el desarrollo, despliegue y uso de estos sistemas inteligentes.</p>
<p>Para ello se realiz&#x00F3; un estudio cualitativo, apoyado en una <italic>metodolog&#x00ED;a</italic> de revisi&#x00F3;n bibliogr&#x00E1;fica de: (i) trabajos acad&#x00E9;micos sobre sus usos actuales y riesgos potenciales de la Inteligencia Artificial; y (ii) un an&#x00E1;lisis comparativo de distintos c&#x00F3;digos &#x00E9;ticos, explorando la convergencia de principios aplicables a Sistemas de Inteligencia Artificial en los contextos educativos.</p>
<p>Los resultados obtenidos, en primer lugar, sit&#x00FA;an los problemas identificados en la tradici&#x00F3;n &#x00E9;tica, cuestionando la proliferaci&#x00F3;n de subdominios de la disciplina. Se indaga, despu&#x00E9;s, la posibilidad de un marco &#x00E9;tico unificado que evite la superposici&#x00F3;n de principios para cada dominio espec&#x00ED;fico. Se constata la utilidad de un marco ampliamente reconocido e influyente, cuyos principios se adaptan bien a los desaf&#x00ED;os de la educaci&#x00F3;n.</p>
<p>Se concluye se&#x00F1;alando las l&#x00ED;neas en las que se debe avanzar en la investigaci&#x00F3;n: (I) fundamentaci&#x00F3;n &#x00E9;tica y regulaci&#x00F3;n normativa para el desarrollo y uso de la Inteligencia Artificial en educaci&#x00F3;n conforme a los principios seleccionados; y (II) definici&#x00F3;n del nuevo perfil profesional docente y sus implicaciones para la formaci&#x00F3;n inicial del profesorado.</p>
</trans-abstract>
<kwd-group xml:lang="en">
<title><italic>Keywords:</italic></title>
<kwd>education</kwd>
<kwd>artificial intelligence</kwd>
<kwd>ethics</kwd>
<kwd>ethical codes</kwd>
<kwd>ethical framework</kwd>
</kwd-group>
<kwd-group xml:lang="es">
<title><italic>Palabras clave:</italic></title>
<kwd>educaci&#x00F3;n</kwd>
<kwd>inteligencia artificial</kwd>
<kwd>&#x00E9;tica</kwd>
<kwd>c&#x00F3;digos &#x00E9;ticos</kwd>
<kwd>marco &#x00E9;tico</kwd>
</kwd-group>
</front-stub>
<body>
<sec id="sec-1-31821-EN">
<label><bold>1.</bold></label>
<title><bold>C<sc>urrent context</sc></bold></title>
<p>The growing digitalisation of everyday tasks in different sectors of human activity has created informational changes that not only modulate our interaction with information and communication technology (ICT) but also how relationships and social processes are articulated. This drives a reontologization of the world rooted in the infosphere and places us in a new era known as hyperhistory (<xref ref-type="bibr" rid="ref-17-31821-EN">Floridi, 2014</xref>). In this era, information is the fundamental resource, meaning that we are vitally dependent on ICT.</p>
<p>Education is what makes this disruptive innovation possible and as such it has now become &#x201C;lifelong&#x201D;. In guaranteeing the transition to information as a resource, the 2030 Agenda establishes that one of the Sustainable Development Goals (SDG 4) is &#x201C;to ensure inclusive and equitable quality education and promote lifelong learning opportunities for all&#x201D; (<xref ref-type="bibr" rid="ref-46-31821-EN">United Nations, 2019</xref>). Artificial Intelligence (AI) can make an effective contribution to these goals as it already offers numerous opportunities in multiple areas of social activity (industry, communication, health, etc.).</p>
<p>Its incorporation in education, driven by <italic>e-learning</italic>, has been discreet, but its use is constantly growing even in the most traditional educational environments. Its potential to improve education is immense, but it is essential that we learn to manage the numerous social, ethical and deontological implications. Above all, because the normalisation of AI in different areas of social life entails an inevitable change to which education is obligated to respond.</p>
<p>According to this &#x201C;discourse of imperative change&#x201D; (<xref ref-type="bibr" rid="ref-6-31821-EN">Bearman <italic>et al.</italic>, 2023</xref>), it is unrealistic to simply dispense with these tools because they entail certain risks. Furthermore, it would be contrary to the ultimate aims of education, which include preparing students to enter the world in which they will have to live &#x2013; a world that is already shaped by digitalisation and AI. It is essential, therefore, to think about <italic>Artificial Intelligence for Education</italic> in the context of <italic>Education for Artificial Intelligence.</italic></p>
<p>Generally speaking, the social impact of AI raises legitimate ethical concerns about its use in educational settings. Being an especially high-risk area&#x2014;and an excellent testing ground for addressing these types of suitably focused problems&#x2014;, it requires serious ethical reflection. This can be approached from an &#x201C;internal&#x201D; perspective, as done by the scientists who create AI, the technologists who implement it and the educators who use it. However, the aim in this study is to focus on an &#x201C;external&#x201D; approach, allowing us to consider the scientific/technological/social three-part dimension of the issues and to suggest criteria for assessing the potential solutions, expected results and potential consequences.</p>
<p>As such, the <italic>objectives</italic> of this study focus on the ethical bases of Artificial Intelligence linked to educational activity, seeking to: a) identify the opportunities, risks and ethical impact associated with its incorporation, which will help understand the need for ethical principles; and b) to specify these principles in order to create an ethical framework to regulate the development, deployment and use of intelligent systems. Other key points to consider include its use in making pedagogical decisions and to determine the skills needed by teachers and students when interacting with AI in order to prevent the associated risks.</p>
<p>To outline a proposal, a qualitative study was performed supported by a literature review based on two types of texts: (i) academic works on uses, risks and ethical guidelines in Artificial Intelligence&#x2014;selected based on the specific topic, authority of the authors and impact of the work&#x2014;that confirmed the need to link AI programs to ethical codes; and (ii) ethical codes for intelligent systems, identifying the convergence of certain principles. Given the amount and diversity of the published codes, an intentional sampling method was selected with the aim of comparing the perspectives of different stakeholders: international organisations, countries involved in the AI race and industries representative of the sector.</p>
<p>In this context, the research follows various steps that set the structure of this article. First, the problem is placed within the philosophical-methodological framework of design science. Second, the uses, opportunities and risks of AI tools in education are described. Third, serious ethical implications are confirmed. In this respect, after confirming continuity with problems in the tradition, the proliferation of subdomains within the discipline is questioned. Fourth, as a result of the analysis of different ethical codes, the convergence of principles that allow for the proposal of a unified framework for AI in education is confirmed. Fifth, the article concludes that future research is required along two different lines: (i) ethical foundation and normative regulation for the development and use of Artificial Intelligence Systems (AIS) in education; and (ii) outlining a new professional teaching profile to identify teacher training needs.</p>
</sec>
<sec id="sec-2-31821-EN">
<label><bold>2.</bold></label>
<title><bold>D<sc>esign science</sc>: <sc>goals</sc>, <sc>processes and results</sc></bold></title>
<p>Artificial Intelligence in Education (AIED) is a scientific undertaking in the realm of the science of design (<xref ref-type="bibr" rid="ref-22-31821-EN">Gonz&#x00E1;lez, 2017a</xref>). Its aim is to create intelligent tools that allow for broader possibilities for different users in specific contexts of use in education. It arises from the confluence of Artificial Intelligence and Education, which are also disciplines of design science as they seek new goals to broaden human possibilities and are articulated in terms of objectives, processes and results.</p>
<p>In contrast to natural and social sciences, design science expressly focuses on predicting how things might be and on prescribing guidelines that help change how things are now to &#x201C;how they should be&#x201D; (<xref ref-type="bibr" rid="ref-41-31821-EN">Simon, 1996</xref>; <xref ref-type="bibr" rid="ref-21-31821-EN">Gonz&#x00E1;lez, 2007</xref>). As such, its main objective is to modulate the future according to a set of predefined aims.</p>
<p>On this scientific basis, interventions can be designed in different fields of human activity and, to put them into practice, they may include technological designs. This is the case with the use of AI to solve problems and improve possibilities in education in terms of achieving objectives. However, this all has serious ethical implications that need to be considered in relation to various aspects:</p>
<p>(i) The objectives, which can be debatable or lead to conflict between different stakeholders. This also includes the selection of problems that AI is used to solve. (ii) The processes: which processes we want to automate, accelerate, etc. and which AI technologies are most suitable for the intended purpose. (iii) The results, which can call into question the intervention itself (that is, the whole sequence of pedagogical activity improved with AI). (iv) The consequences that follow the implemented actions, which may not be as intended.</p>
</sec>
<sec id="sec-3-31821-EN">
<label><bold>3.</bold></label>
<title><bold>U<sc>ses and opportunities of ai in education</sc></bold></title>
<p>AI has a broad range of uses in education (<xref ref-type="bibr" rid="ref-16-31821-EN">Flores Vivar y Garc&#x00ED;a Pe&#x00F1;alvo, 2023</xref>; <xref ref-type="bibr" rid="ref-45-31821-EN">European Union, 2022</xref>; <xref ref-type="bibr" rid="ref-29-31821-EN">Jara y Ochoa, 2020</xref>; <xref ref-type="bibr" rid="ref-37-31821-EN">Oca&#x00F1;a Fern&#x00E1;ndez <italic>et al</italic>., 2019</xref>; <xref ref-type="bibr" rid="ref-34-31821-EN">Moreno Padilla, 2019</xref>; <xref ref-type="bibr" rid="ref-40-31821-EN">S&#x00E1;nchez Vila y Lema Pen&#x00ED;n, 2007</xref>), including the following:</p>
<list list-type="simple">
<list-item><label>(i)</label> <p>To facilitate management tasks, including timetabling, resource allocation, etc.</p></list-item>
<list-item><label>(ii)</label> <p>To automate daily tasks for teachers, including student tracking, sharing information about students with their families or guardians, marking exercises or tests, etc.</p></list-item>
<list-item><label>(iii)</label> <p>To support teaching through intelligent tutoring systems that offer students support depending on the difficulties that they may encounter. These systems can provide students with effective support, even outside the classroom, in the preparation of tasks. In contexts and cases where a human teacher is not available, they can be replaced by <italic>virtual facilitators</italic> (realistic virtual characters based on a combination of AI technology, 3D games and computer animation).</p></list-item>
<list-item><label>(iv)</label> <p>To personalise the learning experience, which is arguably AI&#x2019;s main contribution. The so-called <italic>adaptive teaching systems</italic> are tools related to this objective. They genuinely place the student at the centre of the experience, adjusting the learning path to the individual profile, characteristics and behaviour of each student. By analysing information about each student&#x2019;s progress, they can predict their future performance, allowing for the optimisation of resources and content and even anticipating corrective measures to improve performance.</p></list-item>
</list>
<p>Various AI-based solutions for special educational needs also contribute to a more personalised form of education; for example, live subtitling in the case of hearing impairment or audio description for blind or visually impaired students.</p>
<p>In addition, there are many other algorithmic applications from the world of informal education that are part of the learning ecosystem: through social media, online gaming platforms or mobile apps, among others.</p>
<p>However, although the use of AI in education is becoming the norm, there is still very little empirical research into its real impact. Some studies show that these tools, with a step-based approach to well-defined problems, are successful in supporting teaching and learning in the field of STEM (<xref ref-type="bibr" rid="ref-28-31821-EN">Humble &#x0026; Mozelius, 2022</xref>). And it is precisely these disciplines that seek to develop the right type of skills, from a technological and scientific perspective, to face the challenges of the 21st century (<xref ref-type="bibr" rid="ref-34-31821-EN">Moreno Padilla, 2019</xref>). These AI tools also work well in the learning of foreign languages. In any case, it is important to maintain a critical and supervised attitude (<xref ref-type="bibr" rid="ref-45-31821-EN">European Union, 2022</xref>, p. 14).</p>
</sec>
<sec id="sec-4-31821-EN">
<label><bold>4.</bold></label>
<title><bold>P<sc>otential risks and ethical implications</sc></bold></title>
<p>Limited teacher training in the use of AI is a key concern in terms of its proper implementation in education as this lack of training could lead to the incorrect use or even abuse of the technology. However, to understand the potential risks, it is important to look, above all, at the design of these tools and ask what is behind their introduction in the education &#x201C;market&#x201D;, who is driving it and why.</p>
    <p>The development of AI in education can interfere with people&#x2019;s autonomy and responsibility and obstruct universal rights (<xref ref-type="bibr" rid="ref-5-31821-EN">UN General Assembly, 1948</xref>) such as <italic>privacy</italic> (art. 12), <italic>equality</italic> (art. 1) and <italic>non-discrimination</italic> (art. 2). This has broad social and ethical implications (<xref ref-type="bibr" rid="ref-14-31821-EN">Crawford, 2023</xref>) that also arise in other areas (<xref ref-type="bibr" rid="ref-30-31821-EN">Linares Salgado, 2022</xref>) but which require special consideration in educational contexts.</p>
<p><italic>Privacy</italic>: Machine learning systems are trained using large amounts of data. In education, this means: (i) personal information about students and their families; (ii) records of academic performance; and (iii) tracking data generated from online use and learning activities.</p>
<p>Data is key to the success of AI and forms the basis of personalised education. However, the logic of the technology sector, where everything is seen as data that can be taken and used (<xref ref-type="bibr" rid="ref-14-31821-EN">Crawford, 2023</xref>), poses a series of questions in the education sector about how this data is obtained (consent and privacy), how it is analysed (transparency and trust) and the risk of it being used beyond its approved purpose, where students and families end up as victims of consumer manipulation or other practices (<xref ref-type="bibr" rid="ref-43-31821-EN">Unesco, 2019</xref>).</p>
<p>There is a high risk of cyber-attacks when there are no security protocols. This is especially concerning when minors are involved. Education centres are required to have suitable procedures in place to ensure the protection and ethical use of personal data (<xref ref-type="bibr" rid="ref-45-31821-EN">EU, 2022</xref>, p. 11), but this is no easy task when the skills and knowledge to develop AIED systems are in the hands of for-profit organisations and not within the education sector (<xref ref-type="bibr" rid="ref-28-31821-EN">Humble &#x0026; Mozelius, 2022</xref>).</p>
<p><italic>Equality and non-discrimination</italic>: The data used to train machine learning algorithms is affected by biases from certain contexts and people. These systems then internalise partial or discriminatory criteria (about &#x201C;race&#x201D;, gender, age, etc.) from these sources and can even end up amplifying them.</p>
<p>Historically, different forms of inequality conditioned access to resources and opportunities, and this is now affecting the data used to train algorithms to classify and recognise patterns. As such, &#x201C;under the guise of technical neutrality&#x201D; (<xref ref-type="bibr" rid="ref-14-31821-EN">Crawford, 2023</xref>, p. 231), where groups of users about whom there is less available data often get left behind, AI systems end up perpetuating social inequality. The use of AI is also expanding the digital gap, socially, as the countries and groups of people with the most resources are those who benefit from the opportunities it provides. All this compromises educational fairness.</p>
<p>AI classification practices are also affected by inherent bias. It is the designers who have the final say in deciding what variables are used in the training data sets and which differences should be considered to correctly <italic>classify</italic> new observations (<xref ref-type="bibr" rid="ref-14-31821-EN">Crawford, 2023</xref>). This contributes to maintaining and amplifying stereotypes (with clear epistemological implications that go beyond the aims of this article).</p>
<p><italic>Autonomy:</italic> Interactions with AIS can hinder students&#x2019; development of autonomy and affect their reasoning and decision-making abilities. This adds an additional complication in achieving the goal of creating independent learners (<xref ref-type="bibr" rid="ref-28-31821-EN">Humble &#x0026; Mozelius, 2022</xref>). In addition, relinquishing freedom of choice and delegating decision-making undermines autonomy defined as the capacity for self-legislation and self-determination, linked to the recognition of human dignity.</p>
<p><italic>Responsibility</italic>: AIS are agents &#x2013;and might be replacing human agency&#x2013;, but they are not moral agents. This creates ethical problems related to responsibility (and accountability) for their actions, including cases of legal liability that would need to be addressed in the event of negative consequences for other people. These problems are exacerbated because machine learning algorithms currently present unpredictable elements due to the opacity with which they function.</p>
<sec id="sec-5-31821">
<label>4.1.</label>
<title><italic>Ethical challenges specific to the field of education</italic></title>
<p>Together with these general ethical questions, each of the domains in which AI is used has specific ethical challenges that need to be analysed in context. In education, some of the key considerations are: 1) The potential harm that could arise from educational diagnosis and the prediction of student results that could shape their future development. 2) Problems created by the decisions that an AI system might make due to their impact on the educational decisions of teachers, families and other stakeholders (including legislators). 3) The effect on people&#x2019;s development and maturity, especially in the early years of education, due to a shift in roles that changes the relationship between teachers and students.</p>
<p>It is also important to consider: 4) The pedagogical methods that guide the design of AI systems. Generally speaking, they seek solutions that are most readily monetised, and they are not very innovative. They contribute, therefore, to maintaining the <italic>status quo</italic> of the education system (<xref ref-type="bibr" rid="ref-27-31821-EN">Holmes <italic>et al</italic>, 2022</xref>), losing their immense potential to change education and genuinely broaden people&#x2019;s horizons. 5) The image of the world and the tacit social and political ideas to which AI responds that are transferred to the field of education. This technology is not innocuous and &#x201C;the classifications that are casually chosen to shape a technical system can play a dynamic role in shaping the social and material world&#x201D; (<xref ref-type="bibr" rid="ref-14-31821-EN">Crawford, 2023</xref>, p. 128). 6) It is also important to mention the ethical dimension of some bad practices engaged in by students when using AI tools, which involve fraud and scams, threaten or damage property and compromise intellectual honesty.</p>
</sec>
</sec>
<sec id="sec-6-31821-EN">
<label><bold>5.</bold></label>
<title><bold>T<sc>he importance of ethics</sc></bold></title>
<p>The positive aspects and risks associated with AI are very closely linked, and it is impossible to enjoy the huge benefits that it has to offer without facing its negative consequences. As such, these ambiguities need to be analysed and understood in order to (i) anticipate and calibrate the impacts of certain developments and (ii) select the values and principles that must be protected. This is why the ethics of AI have emerged as a multidisciplinary field of research in applied ethical issues related to the normative problems posed by the development, deployment and use of intelligent systems. These ethics are not for the intelligent systems <italic>per se</italic> but for the people who design, develop or use these systems as they are the ones who could potentially cause, albeit involuntarily, numerous moral problems.</p>
<p>A normative analysis of the impacts of AI reveals parallels with other specific subdomains in the ethics of technology (regarding the internet, data, robots, etc.) formed around different innovations. They often address the same issues (privacy, bias, etc.) with similar approaches, but they are developed in isolation and discussions about them are never connected nor positioned in relation to historical approaches (<xref ref-type="bibr" rid="ref-39-31821-EN">S&#x00E6;tra &#x0026; Danaher, 2022</xref>). This leads to a duplication of work with no increase in actual knowledge and even to the omission of consolidated ideas in the ethics tradition. As such, the difficulties of addressing the ethical challenges of AI often have more to do with forgetting the fundamental questions of ethics &#x2013;and acquired historical knowledge&#x2013; than with our actual understanding of AI.</p>
    <p>The ethics of AI have received a lot of attention from researchers (<xref ref-type="bibr" rid="ref-8-31821-EN">Boddington, 2017</xref>; <xref ref-type="bibr" rid="ref-11-31821-EN">Coeckelberg, 2021</xref>; <xref ref-type="bibr" rid="ref-18-31821-EN">Floridi, 2022</xref>) and from different bodies, institutions and corporations. The current focus is on drawing up codes of ethics to identify the values and principles that should be put into practice. Multiple initiatives and proposals have been put forward in recent years and different countries have also tried to establish suggestions and regulations. Often, the proliferation of codes with their guidelines and principles is seen as a problem (<xref ref-type="bibr" rid="ref-39-31821-EN">S&#x00E6;tra &#x0026; Danaher, 2022</xref>) as it can be confusing (<xref ref-type="bibr" rid="ref-19-31821-EN">Floridi &#x0026; Cowls, 2021</xref>). However, many of these codes have shared principles with certain emphases and nuances based on different cultural and social values.</p>
<p>A comparative analysis of six high-profile ethical codes for AI &#x2013;published between 2017 and 2018<xref ref-type="fn" rid="fn1-EN"><sup>1</sup></xref>&#x2013; allowed <xref ref-type="bibr" rid="ref-19-31821-EN">Floridi &#x0026; Cowls (2021)</xref> to establish a general framework with five core principles, four of which are also used in Bioethics: <italic>beneficence, non-maleficence, autonomy</italic> and <italic>justice</italic> (<xref ref-type="bibr" rid="ref-7-31821-EN">Beauchamp &#x0026; Childress 2019</xref>). However, their transfer to the challenges of AI is not perfect and, in some cases, requires a degree of translation:</p>
<p><italic>Beneficence</italic> refers to promoting wellbeing, preserving dignity and sustaining the planet. <italic>Non-maleficence</italic> refers to privacy, security and caution. <italic>Autonomy</italic> is understood as the power to decide. <italic>Justice</italic> entails promoting prosperity, preserving solidarity and avoiding unfairness. Issues arising from the development of AI itself call for the inclusion of a new principle: <italic>explicability</italic>, which enables other principles through intelligibility and accountability (<xref ref-type="bibr" rid="ref-19-31821-EN">Floridi &#x0026; Cowls, 2021</xref>).</p>
<p>This framework could serve as the architecture within which regulations, technical standards and best practices for specific sectors can be developed. It can play both an enabling role (e.g., the use of AI to achieve the UN Sustainable Development Goals) and a restrictive role (<xref ref-type="bibr" rid="ref-19-31821-EN">Floridi &#x0026; Cowls, 2021</xref>, p. 14). It is a general and realistic ethical framework for monitoring and assessing the design, development and use of AIS because it contains a manageable number of principles that are compatible with universal and irrefutable fundamental values. These principles offer a basis for specific regulations that could arise in a broad range of contexts, for international laws and agreements, the monitoring and assessment mechanisms set up by different countries and even citizen observatories, thereby providing moral, legal and political legitimacy.</p>
</sec>
<sec id="sec-7-31821-EN">
<label><bold>6.</bold></label>
<title><bold>A <sc>unified ethical framework for artificial intelligence in the context of education</sc></bold></title>
<p>Based on the foregoing ideas, we can now move from the general to the specific. This requires adopting ethical frameworks for the use and development of AI in education. The importance of doing so has been highlighted in previous research (<xref ref-type="bibr" rid="ref-27-31821-EN">Holmes <italic>et al</italic>., 2022</xref>; <xref ref-type="bibr" rid="ref-26-31821-EN">Holmes &#x0026; Porayska-Pomsta, 2023</xref>). However, since the <italic>Ethical guidelines for AI in education</italic> published by Aiken &#x0026; Epstein over two decades ago (<xref ref-type="bibr" rid="ref-2-31821-EN">Aiken &#x0026; Epstein 2000</xref>), there has been a surprising lack of published work that focuses specifically on ethics (<xref ref-type="bibr" rid="ref-27-31821-EN">Holmes, <italic>et al.</italic>, 2022</xref>, pp. 505-506). Recently, it has once again become a topic of interest in the education research community (<xref ref-type="bibr" rid="ref-10-31821-EN">Chu <italic>et al</italic>, 2022</xref>; <xref ref-type="bibr" rid="ref-26-31821-EN">Holmes y Porayska-Pomsta, 2023</xref>; <xref ref-type="bibr" rid="ref-9-31821-EN">Boddington, 2023</xref>; <xref ref-type="bibr" rid="ref-35-31821-EN">Nguyen <italic>et al.</italic>, 2023</xref>) and, in the past few years, a number of significant initiatives have been launched by international organisations (<xref ref-type="bibr" rid="ref-43-31821-EN">Unesco, 2019</xref>; <xref ref-type="bibr" rid="ref-45-31821-EN">European Union, 2022</xref>).</p>
<p>The European Union, in its Digital Education Action Plan (2021-2027) &#x2014;which aims to support a sustainable and effective adaptation of the education and training systems of Member States to the digital age&#x2014; established a series of <italic>ethical guidelines on the use of AI and data in teaching and learning for educators</italic> (<xref ref-type="bibr" rid="ref-45-31821-EN">European Union, 2022</xref>). These guidelines, based on the principles and requirements for trustworthy AI (<xref ref-type="bibr" rid="ref-12-31821-EN">European Commission, 2019</xref>), present four key considerations that should guide educators in their decisions about the use of AI: human agency, fairness, humanity and justified choice.</p>
<p>In the academic community, ethical reports and guidelines by international organisations have been systematically analysed and applicable principles selected to ensure that the AI systems developed for education are essentially ethical by design: Principle of governance and stewardship, principle of transparency and accountability, principle of sustainability and proportionality, principle of privacy, principle of security and safety, principle of inclusiveness and principle of human-centered AIED (<xref ref-type="bibr" rid="ref-35-31821-EN">Nguyen <italic>et al</italic>. 2023</xref>).</p>
<p>Considering these proposals, a unified framework is needed to ensure responsible, ethical and trustworthy AI. With this, the aim is to avoid the overlap of principles and recommendations for each specific domain. There is a prestigious framework that has played a key role in the discussions of the European Commission high-level expert group on AI and influenced various recommendations (<xref ref-type="bibr" rid="ref-19-31821-EN">Floridi &#x0026; Cowls, 2021</xref>). Its principles are surprisingly adaptable to a broad range of contexts (<xref ref-type="bibr" rid="ref-30-31821-EN">Linares Salgado, 2022</xref>). The question is whether they can also be applied to issues regarding when, how and why an intelligent system is used in education in order to ensure that AI is used properly in specific teaching decisions, actions and practices.</p>
<p>This question led to a new three-phase analysis:</p>
<list list-type="simple">
<list-item><label>1)</label> <p>Review of guides and recommendations that contain ethical codes deemed relevant due to their impact and published after those analysed by <xref ref-type="bibr" rid="ref-19-31821-EN">Floridi &#x0026; Cowls (2021)</xref>. The aim is to verify the validity and prevalence of the proposed ethical principles. In the sample group, the aim was to ensure that the sector&#x2019;s different influential perspectives were represented: a) recommendations from international organisations: EU (2019), <xref ref-type="bibr" rid="ref-44-31821-EN">UNESCO (2022)</xref> and <xref ref-type="bibr" rid="ref-38-31821-EN">OECD (2019)</xref>; b) documents from countries involved in the AI race: USA (<xref ref-type="bibr" rid="ref-15-31821-EN">Executive Office of the President, 2020</xref>), China (<xref ref-type="bibr" rid="ref-33-31821-EN">China&#x2019;s Ministry of Science and Technology, 2017</xref>), EU (<xref ref-type="bibr" rid="ref-12-31821-EN">EC, 2019</xref>) and Spain (<xref ref-type="bibr" rid="ref-20-31821-EN">Gobierno de Espa&#x00F1;a, 2020</xref>); and c) initiatives from various key companies in the sector: <xref ref-type="bibr" rid="ref-32-31821-EN">Microsoft (2022)</xref>, <xref ref-type="bibr" rid="ref-25-31821-EN">Google (2022)</xref> and Meta (<xref ref-type="bibr" rid="ref-31-31821-EN">Meta Platforms, 2021</xref>).</p></list-item>
<list-item><label>2)</label> <p>Confirmation of the convergence, in the previous ethical framework, of the following aspects: (i) the principles selected by <xref ref-type="bibr" rid="ref-35-31821-EN">Nguyen <italic>et al</italic>. (2023)</xref> and (ii) the European Commission&#x2019;s ethical guidelines for the use of AI in education.</p></list-item>
<list-item><label>3)</label> <p>Verification that the values behind these principles are aligned with the values of education (<xref ref-type="bibr" rid="ref-3-31821-EN">Alonso, 2022</xref> y <xref ref-type="bibr" rid="ref-4-31821-EN">2023</xref>).</p></list-item>
</list>
<sec id="sec-8-31821">
<label>6.1.</label>
<title><italic>Results</italic></title>
<p>The comparative analysis, as seen in <xref ref-type="table" rid="tabw-1-31821-EN">Table 1</xref>, shows a high degree of overlap among the principles of the nine codes, which clearly converge in the general framework proposed by <xref ref-type="bibr" rid="ref-19-31821-EN">Floridi &#x0026; Cowls (2021)</xref>. Although only <italic>explicability</italic> appears in identical form &#x2013;or as <italic>transparency</italic>&#x2013; in all the codes, the &#x201C;translation&#x201D; performed to transfer the principles of bioethics to AI allows for their identification.</p>
<table-wrap id="tabw-1-31821-EN">
<label>T<sc>able</sc> 1</label>
<caption><title>S<sc>ynoptic overview</sc> 1</title></caption>
<table id="tab-1-31821-EN" frame="hsides" border="1" rules="all">
<col width="16%"/>
<col width="17%"/>
<col width="17%"/>
<col width="16%"/>
<col width="17%"/>
<col width="17%"/>
<thead>
<tr>
<th valign="top" align="left"><p>&#x00A0;</p></th>
<th valign="top" align="left"><p><bold>Beneficence</bold></p></th>
<th valign="top" align="left"><p><bold>Non-maleficence</bold></p></th>
<th valign="top" align="left"><p><bold>Autonomy</bold></p></th>
<th valign="top" align="left"><p><bold>Justice</bold></p></th>
<th valign="top" align="left"><p><bold>Explicability</bold></p></th>
</tr>
</thead>
<tbody>
<tr>
<td valign="top" align="left"><p>&#x00A0;</p></td>
<td valign="top" align="left"><p>Promoting wellbeing, preserving dignity and sustaining the planet</p></td>
<td valign="top" align="left"><p>Privacy, security and &#x2018;capability caution&#x2019;</p></td>
<td valign="top" align="left"><p>Power to decide (whether to decide)</p></td>
<td valign="top" align="left"><p>Promoting prosperity, preserving solidarity and avoiding unfairness</p></td>
<td valign="top" align="left"><p>Enabling the other principles through intelligibility and accountability</p></td>
</tr>
<tr>
<td valign="top" align="left"><p><bold>UNESCO</bold></p></td>
<td valign="top" align="left"><p>Sustainability</p></td>
<td valign="top" align="left"><p>Proportionality and do no harm Safety and Security</p></td>
<td valign="top" align="left"><p>Human oversight and determination</p></td>
<td valign="top" align="left"><p>Fairness and non-discrimination</p></td>
<td valign="top" align="left"><p>Transparency and explainability</p></td>
</tr>
<tr>
<td valign="top" align="left"><p><bold>OECD</bold></p></td>
<td valign="top" align="left"><p>Inclusive growth, sustainable development and wellbeing</p></td>
<td valign="top" align="left"><p>Robustness, security and safety</p></td>
<td valign="top" align="left"><p>Respecting human values and fairness (including autonomy)</p></td>
<td valign="top" align="left"><p>Respecting human values and fairness</p></td>
<td valign="top" align="left"><p>Transparency and explicability Accountability</p></td>
</tr>
<tr>
<td valign="top" align="left"><p><bold>EU</bold></p></td>
<td valign="top" align="left"><p>&#x00A0;</p></td>
<td valign="top" align="left"><p>Prevention of harm</p></td>
<td valign="top" align="left"><p>Respect for human autonomy</p></td>
<td valign="top" align="left"><p>Fairness</p></td>
<td valign="top" align="left"><p>Explicability</p></td>
</tr>
<tr>
<td valign="top" align="left"><p><bold>USA</bold></p></td>
<td valign="top" align="left"><p>&#x00A0;</p></td>
<td valign="top" align="left"><p>Safety and Security</p></td>
<td valign="top" align="left"><p>&#x00A0;</p></td>
<td valign="top" align="left"><p>Fairness and non-discrimination</p></td>
<td valign="top" align="left"><p>Disclosure and transparency</p></td>
</tr>
<tr>
<td valign="top" align="left"><p><bold>CHINA</bold></p></td>
<td valign="top" align="left"><p>Harmony and friendship: promoting the common good</p></td>
<td valign="top" align="left"><p>Privacy: respecting and protecting privacy</p></td>
<td valign="top" align="left"><p>Privacy: respecting and protecting people&#x2019;s right to know and choose</p></td>
<td valign="top" align="left"><p>Fairness and justice</p></td>
<td valign="top" align="left"><p>Security: transparency, explainability, traceability, trustworthiness, auditability and safety</p></td>
</tr>
<tr>
<td valign="top" align="left"><p><bold>SPAIN</bold></p></td>
<td valign="top" align="left"><p>Social wellbeing Sustainability</p></td>
<td valign="top" align="left"><p>Data and systems governance: data cannot be used to harm people or to violate their fundamental rights</p></td>
<td valign="top" align="left"><p>Social wellbeing: not reducing, limiting or diverting people&#x2019;s autonomy</p></td>
<td valign="top" align="left"><p>Inclusion</p></td>
<td valign="top" align="left"><p>Transparency</p></td>
</tr>
<tr>
<td valign="top" align="left"><p><bold>GOOGLE</bold></p></td>
<td valign="top" align="left"><p>Be socially beneficial</p></td>
<td valign="top" align="left"><p>Be built and tested for safety Incorporate privacy design principles</p></td>
<td valign="top" align="left"><p>Be accountable to people: the systems must be subject to appropriate human control and direction</p></td>
<td valign="top" align="left"><p>Avoid creating or reinforcing unfair bias</p></td>
<td valign="top" align="left"><p>Be accountable to people: providing appropriate opportunities for feedback, relevant explanations and appeal</p></td>
</tr>
<tr>
<td valign="top" align="left"><p><bold>MICROSOFT</bold></p></td>
<td valign="top" align="left"><p>&#x00A0;</p></td>
<td valign="top" align="left"><p>Privacy and security</p></td>
<td valign="top" align="left"><p>&#x00A0;</p></td>
<td valign="top" align="left"><p>Fairness</p></td>
<td valign="top" align="left"><p>Transparency Accountability</p></td>
</tr>
<tr>
<td valign="top" align="left"><p><bold>META</bold></p></td>
<td valign="top" align="left"><p>&#x00A0;</p></td>
<td valign="top" align="left"><p>Robustness and safety</p></td>
<td valign="top" align="left"><p>&#x00A0;</p></td>
<td valign="top" align="left"><p>Fairness and inclusion</p></td>
<td valign="top" align="left"><p>Transparency and control</p></td>
</tr>
</tbody>
</table>
<table-wrap-foot>
<attrib>Source: prepared by authors</attrib>
</table-wrap-foot>
</table-wrap>
<list list-type="simple">
<list-item><label>I)</label> <p>The principle of <italic>non-maleficence</italic> (which refers back to the principle of <italic>non-laedere</italic> in Roman law)<xref ref-type="fn" rid="fn2-EN"><sup>2</sup></xref> appears expressly as <italic>prevention of harm</italic> in the EU guidelines, as <italic>security and safety</italic> in six codes and is centred around the protection of <italic>privacy</italic> in two others. In Spain&#x2019;s <italic>National AI Strategy</italic>, prevention is associated with the principle of data and systems governance: &#x201C;data cannot be used to harm people or to violate their fundamental rights&#x201D; (<xref ref-type="bibr" rid="ref-20-31821-EN">Gobierno de Espa&#x00F1;a, 2020</xref>, p. 67).</p></list-item>
<list-item><label>II)</label> <p><italic>Justice</italic> <bold>(</bold><italic>suum cuique tribuere</italic>) is referenced explicitly in one of the codes (China), while <italic>fairness</italic> appears more frequently as a criterion of justice (in seven codes). It can be seen in the principle of <italic>inclusion</italic> (Spain) and in the prevention of threats to justice like <italic>avoid creating or reinforcing unfair bias</italic> (Google) and <italic>non-discrimination</italic>.</p></list-item>
<list-item><label>III)</label> <p>The principle of autonomy is associated with human-led decision-making, which could be threatened as the agency of machines increases. Interestingly, this principle is not included in the US <italic>Guidance for the regulation of AI applications</italic> and in two of the three private sector incentives (Microsoft and Meta). It appears in the other codes in the form of <italic>autonomy,</italic> as <italic>human oversight, control or direction</italic> or as the <italic>right to choose.</italic></p></list-item>
<list-item><label>IV)</label> <p>The same codes that do not include the principle of autonomy also do not include the principle of <italic>beneficence</italic>, which limits the creation of AIS to systems that are of benefit to humankind. Nor does it appear explicitly in the ethical guidelines from the EU. However, <italic>social and environmental wellbeing</italic> does appear among the requirements derived from the principle of <italic>fairness</italic>. In the other codes, it appears in various forms linked to <italic>wellbeing</italic> and <italic>sustainability</italic>.</p></list-item>
</list>
<p>In conclusion, the framework presented by <xref ref-type="bibr" rid="ref-19-31821-EN">Floridi &#x0026; Cowls (2021)</xref> offers an ample and complete list of the key ethical principles for AI. &#x2013;As suggested by the authors&#x2013;, it could be a useful guide for regulations and practices in specific domains. It should also include the principles proposed in the field of education, focusing on two different areas: (ii) directing the development, deployment and use of AI tools and (ii) guiding educators in their professional practices.</p>
<p>There are clear parallels between the principles of <italic>non-maleficence</italic> and <italic>beneficence</italic> and the meta-principles proposed by Aiken and Epstein as a core philosophical basis for any discussion about AIED systems. According to the <italic>negative meta-principle</italic>, &#x201C;AIED technology should not diminish the student along any of the fundamental dimensions of human being&#x201D;. The <italic>positive meta-principle</italic> establishes that &#x201C;AIED technology should augment the student along at least one of the fundamental dimensions of human being&#x201D; (<xref ref-type="bibr" rid="ref-2-31821-EN">Aiken &#x0026; Epstein, 2000</xref>, p. 170).</p>
<p>Also, as seen in <xref ref-type="table" rid="tabw-2-31821-EN">Table 2</xref>, there is a high degree of convergence between the proposal from <xref ref-type="bibr" rid="ref-35-31821-EN">Nguyen <italic>et al.</italic> (2023)</xref> and the unified framework: (i) The principle of <italic>non-maleficence</italic> implies <italic>privacy, safety</italic> and <italic>capability caution</italic><bold>.</bold> Therefore, it also incorporates the <italic>principles of privacy and of security and safety</italic> (ii) <italic>Beneficence</italic> refers to <italic>promoting wellbeing, preserving dignity and sustaining the planet</italic>, which relates to the <italic>principle of sustainability and proportionality</italic>. (iii) <italic>Autonomy</italic> means <italic>the power to decide</italic>. Linked to the recognition of people&#x2019;s dignity, it leads to the basic principle that should govern our relationship with AI: <italic>the irreplaceable human element</italic>, as required by the <italic>human-centered AIED principle</italic>. (iv) <italic>Justice</italic> is <italic>promoting prosperity, preserving solidarity and avoiding unfairness</italic>, which is the key aim of the <italic>principle of inclusiveness</italic>. (v) <italic>Explicability</italic> means enabling the other principles through <italic>intelligibility and accountability</italic>. Intelligibility requires <italic>transparency,</italic> and <italic>accountability</italic> implies the <italic>attribution of responsibilities</italic>, as required by the <italic>principle of transparency and accountability</italic>.</p>
<table-wrap id="tabw-2-31821-EN">
<label>T<sc>able</sc> 2</label>
<caption><title>S<sc>ynoptic overview</sc> 2</title></caption>
<table id="tab-2-31821-EN" frame="hsides" border="1" rules="all">
<col width="20%"/>
<col width="20%"/>
<col width="20%"/>
<col width="20%"/>
<col width="20%"/>
<thead>
<tr>
<th valign="top" align="left" rowspan="2" colspan="2"><p><bold><xref ref-type="bibr" rid="ref-19-31821-EN">Floridi &#x0026; Cowls, 2021</xref></bold></p></th>
<th valign="top" align="center" colspan="3"><p><bold>AIED</bold></p></th>
</tr>
<tr>
<th valign="top" align="left"><p><bold><xref ref-type="bibr" rid="ref-35-31821-EN">Nguyen <italic>et al</italic>., 2023</xref></bold></p></th>
<th valign="top" align="left" colspan="2"><p><bold>Education values (<xref ref-type="bibr" rid="ref-45-31821-EN">EU, 2022</xref>)</bold></p></th>
</tr>
<tr>
<th valign="top" align="left"><p><bold>Principles</bold></p></th>
<th valign="top" align="left"><p>&#x00A0;</p></th>
<th valign="top" align="left"><p><bold>Principles</bold></p></th>
<th valign="top" align="left"><p><bold>Ethical considerations</bold></p></th>
<th valign="top" align="left"><p><bold>Values</bold></p></th>
</tr>
</thead>
<tbody>
<tr>
<td valign="top" align="left" rowspan="3"><p><bold>Non-maleficence:</bold></p></td>
<td valign="top" align="left"><p>Privacy</p></td>
<td valign="top" align="left"><p>Privacy</p></td>
<td valign="top" align="left" rowspan="3"><p>&#x00A0;</p></td>
<td valign="top" align="left" rowspan="3"><p>&#x00A0;</p></td>
</tr>
<tr>
<td valign="top" align="left"><p>Safety</p></td>
<td valign="top" align="left"><p>Security and Safety</p></td>
</tr>
<tr>
<td valign="top" align="left"><p>&#x2018;Capacity caution&#x2019;</p></td>
<td valign="top" align="left"><p>&#x00A0;</p></td>
</tr>
<tr>
<td valign="top" align="left" rowspan="4"><p><bold>Beneficence:</bold></p></td>
<td valign="top" align="left" rowspan="2"><p>Promoting wellbeing</p></td>
<td valign="top" align="left" rowspan="2"><p>&#x00A0;</p></td>
<td valign="top" align="left" rowspan="4"><p>Humanity (consideration for people, identity and dignity)</p></td>
<td valign="top" align="left"><p>Wellbeing</p></td>
</tr>
<tr>
<td valign="top" align="left"><p>Safety</p></td>
</tr>
<tr>
<td valign="top" align="left"><p>Preserving dignity</p></td>
<td valign="top" align="left"><p>&#x00A0;</p></td>
<td valign="top" align="left"><p>Social cohesion</p></td>
</tr>
<tr>
<td valign="top" align="left"><p>Sustaining the planet</p></td>
<td valign="top" align="left"><p>Sustainability and proportionality</p></td>
<td valign="top" align="left"><p>Meaningful human connection</p></td>
</tr>
<tr>
<td valign="top" align="left" rowspan="3"><p><bold>Autonomy:</bold></p></td>
<td valign="top" align="left" rowspan="3"><p>Power to decide (to decide) (leading to the irreplaceable human element)</p></td>
<td valign="top" align="left" rowspan="3"><p>Human-centered AIED</p></td>
<td valign="top" align="left" rowspan="3"><p>Human agency</p></td>
<td valign="top" align="left"><p>Autonomy</p></td>
</tr>
<tr>
<td valign="top" align="left"><p>Self-determination</p></td>
</tr>
<tr>
<td valign="top" align="left"><p>Responsibility</p></td>
</tr>
<tr>
<td valign="top" align="left" rowspan="4"><p><bold>Justice:</bold></p></td>
<td valign="top" align="left" rowspan="2"><p>Promoting prosperity</p></td>
<td valign="top" align="left" rowspan="4"><p>Principle of inclusion</p></td>
<td valign="top" align="left" rowspan="4"><p>Fairness</p></td>
<td valign="top" align="left"><p>Equal opportunities</p></td>
</tr>
<tr>
<td valign="top" align="left"><p>Inclusion</p></td>
</tr>
<tr>
<td valign="top" align="left"><p>Preserving solidarity</p></td>
<td valign="top" align="left"><p>Non-discrimination</p></td>
</tr>
<tr>
<td valign="top" align="left"><p>Avoiding unfairness</p></td>
<td valign="top" align="left"><p>Fair distribution of rights and responsibilities</p></td>
</tr>
<tr>
<td valign="top" align="left" rowspan="2"><p><bold>Explicability:</bold></p></td>
<td valign="top" align="left"><p>Intelligibility</p></td>
<td valign="top" align="left" rowspan="2"><p>Principle of transparency and accountability</p></td>
<td valign="top" align="left" rowspan="2"><p>Justified choice</p>
<p>(use of knowledge, facts and data to justify the choices of the various stakeholders in a school setting)</p></td>
<td valign="top" align="left"><p>Transparency</p></td>
</tr>
<tr>
<td valign="top" align="left"><p>Accountability</p>
<p>(Intelligibility requires transparency, and accountability implies the attribution of responsibilities)</p></td>
<td valign="top" align="left"><p>Explicability</p></td>
</tr>
</tbody>
</table>
<table-wrap-foot>
<attrib>Source: prepared by authors</attrib>
</table-wrap-foot>
</table-wrap>
<p>These principles are also aligned with the ethical considerations identified by the European Union to guide educators in their decisions about AI (<xref ref-type="bibr" rid="ref-45-31821-EN">European Union, 2022</xref>, p. 18). The following are recognised in these considerations: (i) the values that form the core aims of education (autonomy, self-determination, responsibility, social cohesion, wellbeing, etc.); (ii) the indispensable values for the development of education (equal opportunities, non-discrimination, transparency, explicability, etc.); and (iii) the values used to inform decision-making in education, including inclusion, fair distribution of rights and responsibilities or meaningful human connection (<xref ref-type="bibr" rid="ref-3-31821-EN">Alonso, 2022</xref> y <xref ref-type="bibr" rid="ref-4-31821-EN">2023</xref>).</p>
<p>The ethical principles of <italic>beneficence, non-maleficence, autonomy, justice</italic> and <italic>explicability</italic> applied to AIS can prevent the risks associated with these systems for them to be used to benefit the Right to Education (SDG 4) as an end in themselves and as a means to achieve the other rights. Any tools that &#x2014;in their design, deployment and use&#x2014; might violate any of these principles would not be legitimate as they would jeopardise the respect for human dignity that human rights seek to preserve.</p>
    <p><italic>Non-maleficence.</italic> AI should be designed to neither aggravate existing harm nor create new harm. As such, it would not be ethical to design tools that hinder the development of any of the inherent facets of human intelligence in its three core domains: perception &#x2014;knowledge&#x2014;, volition and values (<xref ref-type="bibr" rid="ref-23-31821-EN">Gonz&#x00E1;lez, 2017b</xref>), limiting the use of our senses, imagination, thinking and reasoning in a &#x201C;truly human&#x201D; way (<xref ref-type="bibr" rid="ref-36-31821-EN">Nussbaum, 2012</xref>, p. 33).</p>
<p><italic>Beneficence.</italic> In addition to doing no harm, educational AIS should be designed so that students can develop their potential and use all their skills and abilities to choose the life that they want to live in a full and creative manner, in line with human dignity.</p>
<p><italic>Autonomy.</italic> In ethics, autonomy is understood as the ability for self-legislation and self-determination, which is only applicable to human beings. This requires and justifies human control over &#x201C;machines&#x201D;. As such: (a) decisions that affect people&#x2019;s lives cannot be delegated to intelligent machines without human supervision. As a result, educational decisions cannot be solely based on automatic data processing. (b) A human being should always know if they are interacting with human or with a machine. Therefore, when AI tools are used, the special precautions required for vulnerable people, such as children, require informed consent from the parents or legal guardians.</p>
    <p>Education is an especially high-risk area due to its role in the development of human beings and in shaping how we think and act throughout our lives. It is essential to stress caution, which now appears as a requirement of non-maleficence, and to incorporate the <italic>Precautionary Principle</italic> in any educational intervention where AI might present a threat to autonomy and human dignity (expanding on the <xref ref-type="bibr" rid="ref-47-31821-EN">Wingspread Statement</xref>).</p>
</sec>
</sec>
<sec id="sec-9-31821-EN" sec-type="conclusions|discussion">
<label><bold>7.</bold></label>
<title><bold>C<sc>onclusion and discussion</sc></bold></title>
<p>The external perspective on AI and its use in education focuses on social repercussions and responsibility in practices that involved these systems (<xref ref-type="bibr" rid="ref-6-31821-EN">Bearman <italic>et al</italic>., 2023</xref>) and highlights the importance of ethics in the contemporary debate about education (<xref ref-type="bibr" rid="ref-27-31821-EN">Holmes <italic>et al</italic>., 2022</xref>). In this article, a unified framework has been sought for this current and future problem, where science, technology and society have all been taken into account.</p>
<p>This framework is a useful starting point as (i) it can guide and inform coherent and carefully planned regulation, (ii) it allows for the identification of good practices, (iii) it identifies the skills that educators need to develop and (iv) it offers students a set of criteria for interacting with AI and using it to its full potential in learning environments. But there is still a long way to go. There has been a notable increase in research in this field and, in future projects, it would be interesting to compare the parallels detected in this study with the results of a meta-analysis of a much larger group of documents. In turn:</p>
<list list-type="simple">
<list-item><label>1)</label> <p>The principles represent the values that should be achieved through the end goals (e.g., an end goal that, if achieved, could cause harm, would not be legitimate). However, in addition to ensuring ethical goals for AI, we must also determine how they can be achieved in an ethical way because it is one thing to do ethical things and another to do them ethically (<xref ref-type="bibr" rid="ref-27-31821-EN">Holmes <italic>et al</italic>., 2022</xref>, p. 504). As such, even if the end goal is ultimately beneficial, it does not justify the use of all means to achieve it (think about the misuse of data).</p></list-item>
<list-item><label>2)</label> <p>The values to which the principles refer are &#x201C;trusted&#x201D; values, representing some of our core certainties such as the guarantee of human dignity and democratic coexistence. However, (a) they can be conceptualised in different ways, as seen in the different criteria for <italic>justice</italic> used in the different codes and (b) the values are rooted in the Western tradition and, to ensure a true global reach, perspectives from other regions and cultures must be included. This brings us back to the age-old debate of multiculturalism. As argued by Adela Cortina, &#x201C;there can be no response to universal challenges other than to adopt a universal ethical approach, where the end goal of decision-making is the universal good&#x201D; (<xref ref-type="bibr" rid="ref-13-31821-EN">Cortina, 1997</xref>, p. 261).</p></list-item>
<list-item><label>3)</label> <p>The effective mechanisms to ensure that all stakeholders adopt these principles and accept responsibility for their transgression are often not specified. In particular, this refers to the technology companies that develop AI products for education and that also control the current research (<xref ref-type="bibr" rid="ref-1-31821-EN">Ahmed <italic>et al</italic>., 2023</xref>). As social subjects, they can and must take responsibility for their actions insofar as: a) they demonstrate intentionality in the deployed activities and b) decision-making is controlled by the organisation itself (<xref ref-type="bibr" rid="ref-24-31821-EN">Gonz&#x00E1;lez, 2020</xref>).</p></list-item>
</list>
<p>The concerns raised by AI are connected to traditional ethical issues. They are, in general, not new. The degree of significance is different as, to a large extent, AI shapes how we interact with the world, modulates social relationships and, even, affects the configuration of our own identity. The consequences of this affect how we view traditional ideas of morality and individual responsibility, which perhaps need to be reviewed in light of the new conditions.</p>
<p>The debate should develop in two main directions: Firstly, the ethical foundations that give the principles their moral legitimacy. The oversimplification of especially complex issues should be avoided, as can happen when they are addressed by AI experts. To decide what cannot be sacrificed or what must be preferred, we must be able to offer a solid justification of what is considered &#x201C;good&#x201D; and what &#x201C;should be done&#x201D;. And this is a job for ethics.</p>
<p>Ensuring that AIS is developed and used in education in accordance with ethical principles will also require legal and political regulation. However, such laws need to connect with the corresponding values and be in tune with people&#x2019;s reasons and desires. Because having a law (that can be imposed or assumed in the process of socialisation) is one thing and a person&#x2019;s reasons for making it their own is another (<xref ref-type="bibr" rid="ref-13-31821-EN">Cortina, 1997</xref>). As such, an education in values should be considered.</p>
<p>Secondly, the definition of a new professional teaching profile, which requires specifying the expert knowledge that is now needed to work in this profession. This should be included in initial teacher training plans. Based on the above, this should include, at least: (a) scientifically robust research training that combines scientific rigour with practical problem-solving; (b) ethical training to address the challenges of AI; and (c) training in AI that allows educators to adjust their approach and design more effective teaching strategies in order to progress towards the goal of &#x201C;inclusive, fair, quality education and to promote opportunities for lifelong learning for all&#x201D;.</p>
</sec>
</body>
<back>
<fn-group>
<fn id="fn1-EN" fn-type="other"><label><sup>1</sup></label> <p>Asilomar AI Principles (2017); Montreal Declaration for Responsible AI (2017); Version 2 of Ethically Aligned Design (2017); Statement on artificial intelligence, robotics and &#x2018;autonomous&#x2019; systems, from the European Group on Ethics in Science and New Technologies (European Commission, 2018); UK House of Lords Select Committee on Artificial Intelligence Report (2018); Multi-Stakeholder AI Association (2018).</p></fn>
    <fn id="fn2-EN" fn-type="other"><label><sup>2</sup></label> <p>&#x201C;Iuris praecepta sunt haec: honeste vivere, alterum non laedere, suum cuique tribuere&#x201D; (<xref ref-type="bibr" rid="ref-42-31821-EN">Ulpiano</xref>. Digesto 1, 1, 10, 1).</p></fn>
</fn-group>
<ref-list>
<title><bold>R<sc>eferences</sc></bold></title>
<ref id="ref-1-31821-EN"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Ahmed</surname> <given-names>N.</given-names></name> <name><surname>Wahed</surname> <given-names>M.</given-names></name> <name><surname>Thompson</surname> <given-names>N. C.</given-names></name></person-group> <year>2023</year> <article-title>The growing influence of industry in AI research</article-title> <source>Science</source> <volume>379</volume> <issue>6635</issue> <fpage>884</fpage><lpage>886</lpage> <ext-link ext-link-type="uri" xlink:href="https://www.science.org/doi/10.1126/science.ade2420">https://www.science.org/doi/10.1126/science.ade2420</ext-link></element-citation>
<mixed-citation>Ahmed, N., Wahed, M., &#x0026; Thompson, N. C. (2023). The growing influence of industry in AI research. <italic>Science, 379</italic> (6635), 884-886. <ext-link ext-link-type="uri" xlink:href="https://www.science.org/doi/10.1126/">https://www.science.org/doi/10.1126/</ext-link> science.ade2420</mixed-citation></ref>
<ref id="ref-2-31821-EN"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Aiken</surname> <given-names>R. M.</given-names></name> <name><surname>Epstein</surname> <given-names>R. G.</given-names></name></person-group> <year>2000</year> <article-title>Ethical Guidelines for AI in Education: Starting a Conversation</article-title> <source>International Journal of Artificial Intelligence in Education</source> <volume>11</volume> <fpage>163</fpage><lpage>176</lpage></element-citation>
<mixed-citation>Aiken, R. M., &#x0026; Epstein, R. G. (2000). Ethical Guidelines for AI in Education: Starting a Conversation. <italic>International Journal of Artificial Intelligence in Education, 11</italic>, 163-176.</mixed-citation></ref>
<ref id="ref-3-31821-EN"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Alonso</surname> <given-names>A. M.</given-names></name></person-group> <year>2022</year> <article-title>Las Ciencias de Dise&#x00F1;o como Ciencias Aplicadas y el papel de los valores: An&#x00E1;lisis del caso de la Educaci&#x00F3;n</article-title> <source>Revista de Investigaci&#x00F3;n Filos&#x00F3;fica</source> <volume>9</volume><issue>1</issue> <fpage>3</fpage><lpage>27</lpage> <pub-id pub-id-type="doi">10.26754/ojs_arif/arif.202216381</pub-id></element-citation>
<mixed-citation>Alonso, A. M. (2022). Las Ciencias de Dise&#x00F1;o como Ciencias Aplicadas y el papel de los valores: An&#x00E1;lisis del caso de la Educaci&#x00F3;n<italic>. Revista de Investigaci&#x00F3;n Filos&#x00F3;fica, 9</italic>(1), 3-27. <ext-link ext-link-type="uri" xlink:href="https://doi.org/10.26754/ojs_arif/arif.202216381">https://doi.org/10.26754/ojs_arif/arif.202216381</ext-link></mixed-citation></ref>
<ref id="ref-4-31821-EN"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Alonso</surname> <given-names>A. M.</given-names></name></person-group> <year>2023</year> <source>Filosof&#x00ED;a de la Ciencia de la Educaci&#x00F3;n: An&#x00E1;lisis como Ciencia Aplicada de Dise&#x00F1;o</source> <publisher-name>Tirant lo Blanch</publisher-name></element-citation>
<mixed-citation>Alonso, A. M. (2023). <italic>Filosof&#x00ED;a de la Ciencia de la Educaci&#x00F3;n: An&#x00E1;lisis como Ciencia Aplicada de Dise&#x00F1;o</italic>. Tirant lo Blanch.</mixed-citation></ref>
<ref id="ref-5-31821-EN"><element-citation publication-type="webpage"><person-group person-group-type="author"><collab>Asamblea General de la ONU</collab></person-group> <year>1948</year> <source>Declaraci&#x00F3;n Universal de los Derechos Humanos</source> <publisher-loc>Paris</publisher-loc> <ext-link ext-link-type="uri" xlink:href="http://www.un.org/en/universal-declaration-human-rights/">http://www.un.org/en/universal-declaration-human-rights/</ext-link></element-citation>
<mixed-citation>Asamblea General de la ONU. (1948). Declaraci&#x00F3;n Universal de los Derechos Humanos. Paris. <ext-link ext-link-type="uri" xlink:href="http://www.un.org/en/universal-declaration-human-rights/">http://www.un.org/en/universal-declaration-human-rights/</ext-link></mixed-citation></ref>
<ref id="ref-6-31821-EN"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Bearman</surname> <given-names>M.</given-names></name> <name><surname>Ryan</surname> <given-names>J.</given-names></name> <name><surname>Ajjawi</surname> <given-names>R.</given-names></name></person-group> <year>2023</year> <article-title>Discourses of artificial intelligence in higher education: a critical literature review</article-title> <source>Higher Education</source> <volume>86</volume> <fpage>369</fpage><lpage>385</lpage> <pub-id pub-id-type="doi">10.1007/s10734-022-00937-2</pub-id></element-citation>
<mixed-citation>Bearman, M., Ryan, J., &#x0026; Ajjawi, R. (2023). Discourses of artificial intelligence in higher education: a critical literature review. <italic>Higher Education, 86</italic>, 369&#x2013;385. <ext-link ext-link-type="uri" xlink:href="https://doi.org/10.1007/s10734-022-00937-2">https://doi.org/10.1007/s10734-022-00937-2</ext-link></mixed-citation></ref>
<ref id="ref-7-31821-EN"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Beauchamp</surname> <given-names>T. L.</given-names></name> <name><surname>Childress</surname> <given-names>J. F.</given-names></name></person-group> <year>2019</year> <source>Principles of biomedical ethics</source> <publisher-name>Oxford University Press</publisher-name></element-citation>
<mixed-citation>Beauchamp, T. L., &#x0026; Childress, J. F. (2019). <italic>Principles of biomedical ethics</italic>. Oxford University Press.</mixed-citation></ref>
<ref id="ref-8-31821-EN"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Boddington</surname> <given-names>P.</given-names></name></person-group> <year>2017</year> <source>Towards a code of ethics for artificial intelligence</source> <publisher-name>Springer</publisher-name></element-citation>
<mixed-citation>Boddington, P. (2017). <italic>Towards a code of ethics for artificial intelligence</italic>. Springer.</mixed-citation></ref>
<ref id="ref-9-31821-EN"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Boddington</surname> <given-names>P.</given-names></name></person-group> <year>2023</year> <chapter-title>Introduction, Why AI Ethics?</chapter-title> <comment>In</comment> <person-group person-group-type="author"><name><surname>Boddington</surname> <given-names>P.</given-names></name></person-group> <source>AI Ethics. A Textbook</source> <comment>pp.</comment> <fpage>1</fpage><lpage>34</lpage> <publisher-name>Springer</publisher-name></element-citation>
<mixed-citation>Boddington, P. (2023). Introduction, Why AI Ethics? In P. Boddington, <italic>AI Ethics. A Textbook</italic> (pp. 1-34). Springer.</mixed-citation></ref>
<ref id="ref-10-31821-EN"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Chu</surname> <given-names>J.</given-names></name> <name><surname>Xi</surname> <given-names>L.</given-names></name> <name><surname>Zhang</surname> <given-names>Q.</given-names></name> <name><surname>Lin</surname> <given-names>R.</given-names></name></person-group> <year>2022</year> <chapter-title>Research on Ethical Issues of Artificial Intelligence in Education</chapter-title> <comment>In</comment> <person-group person-group-type="editor"><name><surname>Yang</surname> <given-names>J.</given-names></name> <etal>et al.</etal></person-group> <role>Eds.</role> <source>Resilience and Future of Smart Learning</source> <comment>pp.</comment> <fpage>101</fpage><lpage>108</lpage> <publisher-name>Springer</publisher-name> <pub-id pub-id-type="doi">10.1007/978-981-19-5967-7_12</pub-id></element-citation>
<mixed-citation>Chu, J., Xi, L., Zhang, Q., &#x0026; Lin, R. (2022). Research on Ethical Issues of Artificial Intelligence in Education. In J. Yang, <italic>et al.</italic> (Eds.), <italic>Resilience and Future of Smart Learning</italic> (pp. 101-108). Springer. <ext-link ext-link-type="uri" xlink:href="https://doi.org/10.1007/978-981-19-5967-7_12">https://doi.org/10.1007/978-981-19-5967-7_12</ext-link></mixed-citation></ref>
<ref id="ref-11-31821-EN"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Coeckelberg</surname> <given-names>M.</given-names></name></person-group> <year>2021</year> <source>&#x00C9;tica de la Inteligencia Artificial</source> <publisher-name>C&#x00E1;tedra</publisher-name></element-citation>
<mixed-citation>Coeckelberg, M. (2021). <italic>&#x00C9;tica de la Inteligencia Artificial</italic>. C&#x00E1;tedra.</mixed-citation></ref>
<ref id="ref-12-31821-EN"><element-citation publication-type="webpage"><person-group person-group-type="author"><collab>Comisi&#x00F3;n Europea</collab></person-group> <year>2019</year> <article-title>Grupo de Expertos de Alto Nivel sobre IA</article-title> <source>Directrices &#x00C9;ticas para una IA fiable</source> <ext-link ext-link-type="uri" xlink:href="https://op.europa.eu/es/publication-detail/-/publication/d3988569-0434-11ea-8c1f-01aa75ed71a1">https://op.europa.eu/es/publication-detail/-/publication/d3988569-0434-11ea-8c1f-01aa75ed71a1</ext-link></element-citation>
<mixed-citation>Comisi&#x00F3;n Europea. (2019). Grupo de Expertos de Alto Nivel sobre IA. <italic>Directrices &#x00C9;ticas para una IA fiable</italic>. <ext-link ext-link-type="uri" xlink:href="https://op.europa.eu/es/publication-detail/-/publication/d3988569-0434-11ea-8c1f-01aa75ed71a1">https://op.europa.eu/es/publication-detail/-/publication/d3988569-0434-11ea-8c1f-01aa75ed71a1</ext-link></mixed-citation></ref>
<ref id="ref-13-31821-EN"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Cortina</surname> <given-names>A.</given-names></name></person-group> <year>1997</year> <source>Ciudadanos del mundo. Hacia una teor&#x00ED;a de la ciudadan&#x00ED;a</source> <publisher-name>Alianza editorial</publisher-name></element-citation>
<mixed-citation>Cortina, A. (1997). <italic>Ciudadanos del mundo. Hacia una teor&#x00ED;a de la ciudadan&#x00ED;a.</italic> Alianza editorial.</mixed-citation></ref>
<ref id="ref-14-31821-EN"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Crawford</surname> <given-names>K.</given-names></name></person-group> <year>2023</year> <source>Atlas de I. Poder, Pol&#x00ED;tica y Costes Planetarios de la Inteligencia Artificial</source> <publisher-name>Ned ediciones</publisher-name></element-citation>
<mixed-citation>Crawford, K. (2023). <italic>Atlas de I. Poder, Pol&#x00ED;tica y Costes Planetarios de la Inteligencia Artificial</italic>. Ned ediciones.</mixed-citation></ref>
<ref id="ref-15-31821-EN"><element-citation publication-type="webpage"><person-group person-group-type="author"><collab>Executive Office of the President</collab></person-group> <year>2020</year> <source>Guidance for regulation of Artificial Intelligence applications</source> <ext-link ext-link-type="uri" xlink:href="https://www.whitehouse.gov/wp-content/uploads/2020/11/M-21-06.pdf">https://www.whitehouse.gov/wp-content/uploads/2020/11/M-21-06.pdf</ext-link></element-citation>
<mixed-citation>Executive Office of the President. (2020). <italic>Guidance for regulation of Artificial Intelligence applications</italic>. <ext-link ext-link-type="uri" xlink:href="https://www.whitehouse.gov/wp-content/uploads/2020/11/M-21-06.pdf">https://www.whitehouse.gov/wp-content/uploads/2020/11/M-21-06.pdf</ext-link></mixed-citation></ref>
<ref id="ref-16-31821-EN"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Flores Vivar</surname> <given-names>J. M.</given-names></name> <name><surname>Garc&#x00ED;a-Pe&#x00F1;alvo</surname> <given-names>F. J.</given-names></name></person-group> <year>2023</year> <article-title>Reflexiones sobre la &#x00E9;tica, potencialidades y retos de la Inteligencia Artificial en el marco de la Educaci&#x00F3;n de calidad (ODS4)</article-title> <source>Comunicar</source> <volume>31</volume><issue>74</issue> <fpage>37</fpage><lpage>47</lpage> <pub-id pub-id-type="doi">10.3916/C74-2023-03</pub-id></element-citation>
<mixed-citation>Flores Vivar, J. M. y Garc&#x00ED;a-Pe&#x00F1;alvo, F. J. (2023). Reflexiones sobre la &#x00E9;tica, potencialidades y retos de la Inteligencia Artificial en el marco de la Educaci&#x00F3;n de calidad (ODS4). <italic>Comunicar, 31</italic>(74), 37-47. <ext-link ext-link-type="uri" xlink:href="https://doi.org/10.3916/C74-2023-03">https://doi.org/10.3916/C74-2023-03</ext-link></mixed-citation></ref>
<ref id="ref-17-31821-EN"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Floridi</surname> <given-names>L.</given-names></name></person-group> <year>2014</year> <source>The Fourth Revolution - How the Infosphere is Reshaping Human Reality</source> <publisher-name>Oxford University Press</publisher-name></element-citation>
<mixed-citation>Floridi, L. (2014). <italic>The Fourth Revolution - How the Infosphere is Reshaping Human Reality.</italic> Oxford University Press.</mixed-citation></ref>
<ref id="ref-18-31821-EN"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Floridi</surname> <given-names>L</given-names></name></person-group> <year>2022</year> <source>Etica dell&#x2019;intelligenza artificiale. Sviluppi, opportunit&#x00E0;, sfide</source> <publisher-name>Rafaello Cortina</publisher-name></element-citation>
<mixed-citation>Floridi, L (2022). <italic>Etica dell&#x2019;intelligenza artificiale. Sviluppi, opportunit&#x00E0;, sfide.</italic> Rafaello Cortina.</mixed-citation></ref>
<ref id="ref-19-31821-EN"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Floridi</surname> <given-names>L.</given-names></name> <name><surname>Cowls</surname> <given-names>J.</given-names></name></person-group> <year>2021</year> <chapter-title>A Unified Framework of Five Principles for AI in Society</chapter-title> <comment>In</comment> <person-group person-group-type="editor"><name><surname>Floridi</surname> <given-names>L.</given-names></name></person-group> <role>Ed.</role> <source>Ethics, Governance, and Policies in Artificial Intelligence</source> <fpage>5</fpage><lpage>17</lpage> <publisher-name>Springer</publisher-name></element-citation>
<mixed-citation>Floridi L., &#x0026; Cowls, J. (2021). A Unified Framework of Five Principles for AI in Society. In L. Floridi (Ed.), <italic>Ethics, Governance, and Policies in Artificial Intelligence</italic> (5-17). Springer.</mixed-citation></ref>
<ref id="ref-20-31821-EN"><element-citation publication-type="webpage"><person-group person-group-type="author"><collab>Gobierno de Espa&#x00F1;a</collab></person-group> <year>2020</year> <source>Estrategia Nacional de Inteligencia Artificial</source> <ext-link ext-link-type="uri" xlink:href="https://www.lamoncloa.gob.es/presidente/actividades/Documents/2020/021220-ENIA.pdf">https://www.lamoncloa.gob.es/presidente/actividades/Documents/2020/021220-ENIA.pdf</ext-link></element-citation>
<mixed-citation>Gobierno de Espa&#x00F1;a. (2020). <italic>Estrategia Nacional de Inteligencia Artificial</italic>. <ext-link ext-link-type="uri" xlink:href="https://www.lamoncloa.gob.es/presidente/actividades/Documents/2020/021220-ENIA.pdf">https://www.lamoncloa.gob.es/presidente/actividades/Documents/2020/021220-ENIA.pdf</ext-link></mixed-citation></ref>
<ref id="ref-21-31821-EN"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Gonz&#x00E1;lez</surname> <given-names>W. J.</given-names></name></person-group> <year>2007</year> <chapter-title>Configuraci&#x00F3;n de las Ciencias de Dise&#x00F1;o como Ciencias de lo Artificial: Papel de la Inteligencia Artificial y de la racionalidad limitada</chapter-title> <comment>En</comment> <person-group person-group-type="editor"><name><surname>Gonz&#x00E1;lez</surname> <given-names>W. J.</given-names></name></person-group> <role>Ed.</role> <source>Las Ciencias de Dise&#x00F1;o: Racionalidad limitada, predicci&#x00F3;n y prescripci&#x00F3;n</source> <fpage>41</fpage><lpage>69</lpage> <publisher-name>Netbiblo</publisher-name></element-citation>
<mixed-citation>Gonz&#x00E1;lez, W. J. (2007). Configuraci&#x00F3;n de las Ciencias de Dise&#x00F1;o como Ciencias de lo Artificial: Papel de la Inteligencia Artificial y de la racionalidad limitada. En W. J. Gonz&#x00E1;lez (Ed.), <italic>Las Ciencias de Dise&#x00F1;o: Racionalidad limitada, predicci&#x00F3;n y prescripci&#x00F3;n</italic> (41-69). Netbiblo.</mixed-citation></ref>
<ref id="ref-22-31821-EN"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Gonz&#x00E1;lez</surname> <given-names>W. J.</given-names></name></person-group> <year>2017a</year> <article-title>Artificial Intelligence in a New Context: &#x201C;Internal&#x201D; and &#x201C;External&#x201D; Factors</article-title> <source>Minds &#x0026; Machines</source> <volume>27</volume><issue>3</issue> <fpage>393</fpage><lpage>396</lpage> <pub-id pub-id-type="doi">10.1007/s11023-017-9444-3</pub-id></element-citation>
<mixed-citation>Gonz&#x00E1;lez, W. J. (2017a). Artificial Intelligence in a New Context: &#x201C;Internal&#x201D; and &#x201C;External&#x201D; Factors. <italic>Minds &#x0026; Machines, 27</italic>(3), 393&#x2013;396. <ext-link ext-link-type="uri" xlink:href="https://doi.org/10.1007/s11023-017-9444-3">https://doi.org/10.1007/s11023-017-9444-3</ext-link></mixed-citation></ref>
<ref id="ref-23-31821-EN"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Gonz&#x00E1;lez</surname> <given-names>W. J.</given-names></name></person-group> <year>2017b</year> <article-title>From Intelligence to Rationality of Minds and Machines in Contemporary Society: The Sciences of Design and the Role of Information</article-title> <source>Minds &#x0026; Machines</source> <volume>27</volume><issue>3</issue> <fpage>397</fpage><lpage>424</lpage> <pub-id pub-id-type="doi">10.1007/s11023-017-9439-0</pub-id></element-citation>
<mixed-citation>Gonz&#x00E1;lez, W. J. (2017b). From Intelligence to Rationality of Minds and Machines in Contemporary Society: The Sciences of Design and the Role of Information. <italic>Minds &#x0026; Machines, 27</italic>(3), 397-424. <ext-link ext-link-type="uri" xlink:href="https://doi.org/10.1007/s11023-017-9439-0">https://doi.org/10.1007/s11023-017-9439-0</ext-link></mixed-citation></ref>
<ref id="ref-24-31821-EN"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Gonz&#x00E1;lez</surname> <given-names>W. J.</given-names></name></person-group> <year>2020</year> <article-title>The Internet at the Service of Society: Business Ethics, Rationality, and Responsibility</article-title> <source>&#x00C9;ndoxa</source> <volume>46</volume> <fpage>383</fpage><lpage>412</lpage> <pub-id pub-id-type="doi">10.5944/endoxa.46.2020.28029</pub-id></element-citation>
<mixed-citation>Gonz&#x00E1;lez, W. J. (2020). The Internet at the Service of Society: Business Ethics, Rationality, and Responsibility. <italic>&#x00C9;ndoxa, 46</italic>, 383-412. <ext-link ext-link-type="uri" xlink:href="https://doi.org/10.5944/endoxa.46.2020.28029">https://doi.org/10.5944/endoxa.46.2020.28029</ext-link></mixed-citation></ref>
<ref id="ref-25-31821-EN"><element-citation publication-type="webpage"><person-group person-group-type="author"><collab>Google LLC</collab></person-group> <year>2022</year> <source>Artificial Intelligence at Google: Our Principles</source> <ext-link ext-link-type="uri" xlink:href="https://ai.google/principles/">https://ai.google/principles/</ext-link></element-citation>
<mixed-citation>Google LLC. (2022). <italic>Artificial Intelligence at Google: Our Principles</italic>. <ext-link ext-link-type="uri" xlink:href="https://ai.google/principles/">https://ai.google/principles/</ext-link></mixed-citation></ref>
<ref id="ref-26-31821-EN"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Holmes</surname> <given-names>W.</given-names></name> <name><surname>Porayska-Pomsta</surname> <given-names>K.</given-names></name></person-group> <year>2023</year> <source>The Ethics of Artificial Intelligence in Education</source> <publisher-name>Routledge Taylor</publisher-name></element-citation>
<mixed-citation>Holmes, W., &#x0026; Porayska-Pomsta, K. (2023). <italic>The Ethics of Artificial Intelligence in Education.</italic> Routledge Taylor.</mixed-citation></ref>
<ref id="ref-27-31821-EN"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Holmes</surname> <given-names>W.</given-names></name> <name><surname>Porayska-Pomsta</surname> <given-names>K.</given-names></name> <name><surname>Holstein</surname> <given-names>K.</given-names></name> <etal>et al.</etal></person-group> <year>2022</year> <article-title>Ethics of AI in Education: Towards a Community-Wide Framework</article-title> <source>International Journal of Artificial Intelligence in Education</source> <volume>32</volume> <fpage>504</fpage><lpage>526</lpage> <pub-id pub-id-type="doi">10.1007/s40593-021-00239-1</pub-id></element-citation>
<mixed-citation>Holmes, W., Porayska-Pomsta, K., Holstein, K. et al. (2022). Ethics of AI in Education: Towards a Community-Wide Framework. <italic>International Journal of Artificial Intelligence in Education, 32</italic>, 504&#x2013;526. <ext-link ext-link-type="uri" xlink:href="https://doi.org/10.1007/s40593-021-00239-1">https://doi.org/10.1007/s40593-021-00239-1</ext-link></mixed-citation></ref>
<ref id="ref-28-31821-EN"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Humble</surname> <given-names>N.</given-names></name> <name><surname>Mozellius</surname> <given-names>P.</given-names></name></person-group> <year>2022</year> <article-title>The threat, hype, and promise of artificial intelligence in education</article-title> <source>Discover Artificial Intelligence</source> <volume>2</volume><issue>22</issue> <pub-id pub-id-type="doi">10.1007/s44163-022-00039-z</pub-id></element-citation>
<mixed-citation>Humble, N., &#x0026; Mozellius, P. (2022). The threat, hype, and promise of artificial intelligence in education. <italic>Discover Artificial Intelligence, 2</italic>(22). <ext-link ext-link-type="uri" xlink:href="https://doi.org/10.1007/s44163-022-00039-z">https://doi.org/10.1007/s44163-022-00039-z</ext-link></mixed-citation></ref>
<ref id="ref-29-31821-EN"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Jara</surname> <given-names>I.</given-names></name> <name><surname>Ochoa</surname> <given-names>J. M.</given-names></name></person-group> <year>2020</year> <source>Usos y efectos de la inteligencia artificial en educaci&#x00F3;n</source> <publisher-name>Banco Interamericano de Desarrollo</publisher-name> <pub-id pub-id-type="doi">10.18235/0002380</pub-id></element-citation>
<mixed-citation>Jara, I., &#x0026; Ochoa, J. M. (2020). <italic>Usos y efectos de la inteligencia artificial en educaci&#x00F3;n</italic>. Banco Interamericano de Desarrollo. <ext-link ext-link-type="uri" xlink:href="http://dx.doi.org/10.18235/0002380">http://dx.doi.org/10.18235/0002380</ext-link></mixed-citation></ref>
<ref id="ref-30-31821-EN"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Linares Salgado</surname> <given-names>J. E.</given-names></name></person-group> <year>2022</year> <article-title>Principios &#x00E9;ticos para el desarrollo de la Inteligencia Artificial y su aplicaci&#x00F3;n en los sistemas de salud</article-title> <source>ArtefaCToS. Revista de Estudios de la Ciencia y la Tecnolog&#x00ED;a</source> <volume>11</volume><issue>2</issue> <fpage>137</fpage><lpage>161</lpage></element-citation>
<mixed-citation>Linares Salgado, J. E. (2022). Principios &#x00E9;ticos para el desarrollo de la Inteligencia Artificial y su aplicaci&#x00F3;n en los sistemas de salud. <italic>ArtefaCToS. Revista de Estudios de la Ciencia y la Tecnolog&#x00ED;a, 11</italic>(2), 137-161.</mixed-citation></ref>
<ref id="ref-31-31821-EN"><element-citation publication-type="webpage"><person-group person-group-type="author"><collab>Meta Platforms</collab></person-group> <year>2021</year> <source>Facebook five pillars of responsible AI</source> <ext-link ext-link-type="uri" xlink:href="https://ai.facebook.com/blog/facebooks-five-pillars-of-responsible-ai/">https://ai.facebook.com/blog/facebooks-five-pillars-of-responsible-ai/</ext-link></element-citation>
<mixed-citation>Meta Platforms. (2021). <italic>Facebook five pillars of responsible AI</italic>. <ext-link ext-link-type="uri" xlink:href="https://ai.facebook.com/blog/facebooks-five-pillars-of-responsible-ai/">https://ai.facebook.com/blog/facebooks-five-pillars-of-responsible-ai/</ext-link></mixed-citation></ref>
<ref id="ref-32-31821-EN"><element-citation publication-type="webpage"><person-group person-group-type="author"><collab>Microsoft Corporation</collab></person-group> <year>2022</year> <source>Responsible IA</source> <ext-link ext-link-type="uri" xlink:href="https://www.microsoft.com/en-us/ai/responsible-ai?activetab=pivot1%3aprimaryr6">https://www.microsoft.com/en-us/ai/responsible-ai?activetab=pivot1%3aprimaryr6</ext-link></element-citation>
<mixed-citation>Microsoft Corporation (2022). <italic>Responsible IA</italic>. <ext-link ext-link-type="uri" xlink:href="https://www.microsoft.com/en-us/ai/responsible-ai?activetab=pivot1%3aprimaryr6">https://www.microsoft.com/en-us/ai/responsible-ai?activetab=pivot1%3aprimaryr6</ext-link></mixed-citation></ref>
<ref id="ref-33-31821-EN"><element-citation publication-type="webpage"><person-group person-group-type="author"><collab>Ministerio de Ciencia y Tecnolog&#x00ED;a de la Rep&#x00FA;blica Popular China</collab></person-group> <year>2017</year> <source>Principios de Gobernanza para una Nueva Generaci&#x00F3;n de IA: Desarrollo Responsable de la Inteligencia Artificial</source> <ext-link ext-link-type="uri" xlink:href="https://www.newamerica.org/cybersecurity-initiative/digichina/blog/translation-chinese-expert-group-offers-governance-principles-responsible-ai/">https://www.newamerica.org/cybersecurity-initiative/digichina/blog/translation-chinese-expert-group-offers-governance-principles-responsible-ai/</ext-link></element-citation>
<mixed-citation>Ministerio de Ciencia y Tecnolog&#x00ED;a de la Rep&#x00FA;blica Popular China. (2017). <italic>Principios de Gobernanza para una Nueva Generaci&#x00F3;n de IA: Desarrollo Responsable de la Inteligencia Artificial</italic>. <ext-link ext-link-type="uri" xlink:href="https://www.newamerica.org/cybersecurity-initiative/digichina/blog/translation-chinese-expert-group-offers-governance-principles-responsible-ai/">https://www.newamerica.org/cybersecurity-initiative/digichina/blog/translation-chinese-expert-group-offers-governance-principles-responsible-ai/</ext-link></mixed-citation></ref>
<ref id="ref-34-31821-EN"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Moreno Padilla</surname> <given-names>R.</given-names></name></person-group> <year>2019</year> <article-title>La llegada de la inteligencia artificial a la educaci&#x00F3;n</article-title> <source>RITI Journal</source> <volume>7</volume><issue>14</issue> <fpage>260</fpage><lpage>270</lpage> <pub-id pub-id-type="doi">10.36825/RITI.07.14.022</pub-id></element-citation>
<mixed-citation>Moreno Padilla, R. (2019). La llegada de la inteligencia artificial a la educaci&#x00F3;n. <italic>RITI Journal, 7</italic>(14), 260-270. <ext-link ext-link-type="uri" xlink:href="https://doi.org/10.36825/RITI.07.14.022">https://doi.org/10.36825/RITI.07.14.022</ext-link></mixed-citation></ref>
<ref id="ref-35-31821-EN"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Nguyen</surname> <given-names>A</given-names></name> <name><surname>Ngo</surname> <given-names>H. N</given-names></name> <name><surname>Hong</surname> <given-names>Y.</given-names></name> <name><surname>Dang</surname> <given-names>B.</given-names></name> <name><surname>Nguyen</surname> <given-names>B.T.</given-names></name></person-group> <year>2023</year> <article-title>Ethical principles for artificial intelligence in education</article-title> <source>Educ Inf Technol</source> <volume>28</volume><issue>4</issue> <fpage>4221</fpage><lpage>4241</lpage> <pub-id pub-id-type="doi">10.1007/s10639-022-11316-w</pub-id></element-citation>
<mixed-citation>Nguyen A, Ngo H. N, Hong, Y., Dang, B., &#x0026; Nguyen, B.T. (2023). Ethical principles for artificial intelligence in education. <italic>Educ Inf Technol, 28</italic>(4), 4221-4241. <ext-link ext-link-type="uri" xlink:href="https://doi.org/10.1007/s10639-022-11316-w">https://doi.org/10.1007/s10639-022-11316-w</ext-link></mixed-citation></ref>
<ref id="ref-36-31821-EN"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Nussbaum</surname> <given-names>M.</given-names></name></person-group> <year>2012</year> <source>Crear capacidades. Propuesta para el desarrollo humano</source> <publisher-name>Paid&#x00F3;s</publisher-name></element-citation>
<mixed-citation>Nussbaum, M. (2012). <italic>Crear capacidades. Propuesta para el desarrollo humano.</italic> Paid&#x00F3;s.</mixed-citation></ref>
<ref id="ref-37-31821-EN"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Oca&#x00F1;a-Fernandez</surname> <given-names>Y.</given-names></name> <name><surname>Valenzuela-Fernandez</surname> <given-names>L.</given-names></name> <name><surname>Garro-Aburto</surname> <given-names>L.</given-names></name></person-group> <year>2019</year> <article-title>Inteligencia artificial y sus implicaciones en la educaci&#x00F3;n superior</article-title> <source>Prop&#x00F3;sitos y Representaciones</source> <volume>7</volume><issue>2</issue> <fpage>536</fpage><lpage>552</lpage> <pub-id pub-id-type="doi">10.20511/pyr2019.v7n2.274</pub-id></element-citation>
<mixed-citation>Oca&#x00F1;a-Fernandez, Y., Valenzuela-Fernandez, L., &#x0026; Garro-Aburto, L. (2019). Inteligencia artificial y sus implicaciones en la educaci&#x00F3;n superior. <italic>Prop&#x00F3;sitos y Representaciones, 7</italic>(2), 536-552. <ext-link ext-link-type="uri" xlink:href="http://dx.doi.org/10.20511/pyr2019.v7n2.274">http://dx.doi.org/10.20511/pyr2019.v7n2.274</ext-link></mixed-citation></ref>
<ref id="ref-38-31821-EN"><element-citation publication-type="webpage"><person-group person-group-type="author"><collab>OECD</collab></person-group> <year>2019</year> <source>Recommendation of the Council on Artificial Intelligence</source> <comment>OECD/LEGAL/0449</comment> <ext-link ext-link-type="uri" xlink:href="https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0449">https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0449</ext-link></element-citation>
<mixed-citation>OECD (2019). <italic>Recommendation of the Council on Artificial Intelligence</italic>. OECD/LEGAL/0449. <ext-link ext-link-type="uri" xlink:href="https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0449">https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0449</ext-link></mixed-citation></ref>
<ref id="ref-39-31821-EN"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>S&#x00E6;tra</surname> <given-names>H.S.</given-names></name> <name><surname>Danaher</surname> <given-names>J.</given-names></name></person-group> <year>2022</year> <article-title>To Each Technology Its Own Ethics: The Problem of Ethical Proliferation</article-title> <source>Philosophy &#x0026; Technology</source> <volume>35</volume> <fpage>93</fpage> <pub-id pub-id-type="doi">10.1007/s13347-022-00591-7</pub-id></element-citation>
<mixed-citation>S&#x00E6;tra, H.S., &#x0026; Danaher, J. (2022). To Each Technology Its Own Ethics: The Problem of Ethical Proliferation. <italic>Philosophy &#x0026; Technology, 35</italic>, 93. <ext-link ext-link-type="uri" xlink:href="https://doi.org/10.1007/s13347-022-00591-7">https://doi.org/10.1007/s13347-022-00591-7</ext-link></mixed-citation></ref>
<ref id="ref-40-31821-EN"><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>S&#x00E1;nchez Vila</surname> <given-names>E. M.</given-names></name> <name><surname>Lama Pen&#x00ED;n</surname> <given-names>M.</given-names></name></person-group> <year>2007</year> <article-title>Monograf&#x00ED;a: T&#x00E9;cnicas de la Inteligencia Artificial Aplicadas a la Educaci&#x00F3;n</article-title> <source>Inteligencia Artificial. Revista Iberoamericana de Inteligencia Artificial</source> <volume>11</volume><issue>33</issue> <fpage>7</fpage><lpage>12</lpage></element-citation>
<mixed-citation>S&#x00E1;nchez Vila, E. M., &#x0026; Lama Pen&#x00ED;n, M. (2007). Monograf&#x00ED;a: T&#x00E9;cnicas de la Inteligencia Artificial Aplicadas a la Educaci&#x00F3;n. <italic>Inteligencia Artificial. Revista Iberoamericana de Inteligencia Artificial, 11</italic>(33), 7-12.</mixed-citation></ref>
<ref id="ref-41-31821-EN"><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Simon</surname> <given-names>H. A.</given-names></name></person-group> <year>1996</year> <source>The Sciences of the Artificial</source> <publisher-name>The MIT Press</publisher-name> <edition>3&#x00AA; ed.</edition></element-citation>
<mixed-citation>Simon, H. A. (1996). <italic>The Sciences of the Artificial</italic>. The MIT Press. 3&#x00AA; ed.</mixed-citation></ref>
<ref id="ref-42-31821-EN"><element-citation publication-type="webpage"><person-group person-group-type="author"><name><surname>Ulpiano</surname> <given-names>E. D.</given-names></name></person-group> <year>1836</year> <article-title>Digesto</article-title> <comment>En</comment> <person-group person-group-type="author"><collab>Academicun parisiense</collab></person-group> <source>Corpus Iuris civilis</source> <ext-link ext-link-type="uri" xlink:href="https://gallica.bnf.fr/ark:/12148/bpt6k65768518/f13.item.r=digestum">https://gallica.bnf.fr/ark:/12148/bpt6k65768518/f13.item.r=digestum</ext-link></element-citation>
<mixed-citation>Ulpiano, E. D. (1836). Digesto. En Academicun parisiense, <italic>Corpus Iuris civilis</italic>. <ext-link ext-link-type="uri" xlink:href="https://gallica.bnf.fr/ark:/12148/bpt6k65768518/f13.item.r=digestum">https://gallica.bnf.fr/ark:/12148/bpt6k65768518/f13.item.r=digestum</ext-link></mixed-citation></ref>
<ref id="ref-43-31821-EN"><element-citation publication-type="webpage"><person-group person-group-type="author"><collab>Unesco</collab></person-group> <year>2019</year> <source>Beijing consensus on artificial intelligence and education</source> <ext-link ext-link-type="uri" xlink:href="https://unesdoc.unesco.org/ask:/48223/pf0000368303">https://unesdoc.unesco.org/ask:/48223/pf0000368303</ext-link></element-citation>
<mixed-citation>Unesco. (2019). <italic>Beijing consensus on artificial intelligence and education</italic>. <ext-link ext-link-type="uri" xlink:href="https://unesdoc.unesco.org/ask:/48223/pf0000368303">https://unesdoc.unesco.org/ask:/48223/pf0000368303</ext-link></mixed-citation></ref>
<ref id="ref-44-31821-EN"><element-citation publication-type="webpage"><person-group person-group-type="author"><collab>Unesco</collab></person-group> <year>2022</year> <source>Recomendaci&#x00F3;n sobre la &#x00E9;tica de la inteligencia artificial</source> <ext-link ext-link-type="uri" xlink:href="https://unesdoc.unesco.org/ark:/48223/pf0000379920_spa.page=15">https://unesdoc.unesco.org/ark:/48223/pf0000379920_spa.page=15</ext-link></element-citation>
<mixed-citation>Unesco. (2022). <italic>Recomendaci&#x00F3;n sobre la &#x00E9;tica de la inteligencia artificial</italic>. <ext-link ext-link-type="uri" xlink:href="https://unesdoc.unesco.org/ark:/48223/pf0000379920_spa.page=15">https://unesdoc.unesco.org/ark:/48223/pf0000379920_spa.page=15</ext-link>.</mixed-citation></ref>
<ref id="ref-45-31821-EN"><element-citation publication-type="webpage"><person-group person-group-type="author"><collab>Uni&#x00F3;n Europea</collab></person-group> <year>2022</year> <source>Directrices &#x00E9;ticas sobre el uso de la inteligencia artificial (IA) y los datos en la educaci&#x00F3;n y formaci&#x00F3;n para los educadores</source> <publisher-name>Publications Office of the European Union</publisher-name> <ext-link ext-link-type="uri" xlink:href="https://data.europa.eu/doi/10.2766/898">https://data.europa.eu/doi/10.2766/898</ext-link></element-citation>
<mixed-citation>Uni&#x00F3;n Europea. (2022). <italic>Directrices &#x00E9;ticas sobre el uso de la inteligencia artificial (IA) y los datos en la educaci&#x00F3;n y formaci&#x00F3;n para los educadores</italic>. Publications Office of the European Union. <ext-link ext-link-type="uri" xlink:href="https://data.europa.eu/doi/10.2766/898">https://data.europa.eu/doi/10.2766/898</ext-link></mixed-citation></ref>
<ref id="ref-46-31821-EN"><element-citation publication-type="report"><person-group person-group-type="author"><collab>United Nations</collab></person-group> <year>2019</year> <source>The Sustainable Development Goals Report</source> <volume>2019</volume> <ext-link ext-link-type="uri" xlink:href="https://bit.ly/34nbq60">https://bit.ly/34nbq60</ext-link></element-citation>
<mixed-citation>United Nations. (2019). <italic>The Sustainable Development Goals Report 2019</italic>. <ext-link ext-link-type="uri" xlink:href="https://bit.ly/34nbq60">https://bit.ly/34nbq60</ext-link></mixed-citation></ref>
<ref id="ref-47-31821-EN"><element-citation publication-type="webpage"><article-title>Wingspread Statement on the Precautionary Principle</article-title> <year>1998</year> <ext-link ext-link-type="uri" xlink:href="http://www.sehn.org/sehn/the-precautionary-principle-march-1998">http://www.sehn.org/sehn/the-precautionary-principle-march-1998</ext-link></element-citation>
<mixed-citation>Wingspread Statement on the Precautionary Principle (1998). <ext-link ext-link-type="uri" xlink:href="http://www.sehn.org/sehn/the-precautionary-principle-march-1998">http://www.sehn.org/sehn/the-precautionary-principle-march-1998</ext-link></mixed-citation></ref>
</ref-list>
</back>
</sub-article>
</article>
