The right to education under Article 2 of Protocol No. 1 to the ECHR, taken in conjunction with Article 14 ECHR,  require that States address risks related to non-discrimination and equality for all individuals in educational contexts.[1] In the context of AI systems, it entails ensuring that AI systems do not reinforce biases which may lead to discriminatory outcomes or create barriers to access to education, or inequalities in the standard and quality of education. Children, in particular, due to their stage of development, have specific needs and rights that distinguish them from adults. As such, there is a need for child-focused approach in the procurement and use of educational technology, including AI systems.[2] In all actions concerning children, whether undertaken by public or private social welfare institutions, courts of law, administrative authorities or legislative bodies, the best interests of the child shall be a primary consideration.[3]

Given the increasing importance of new technologies, a number of Council of Europe documents have been adopted in this area, which invite states to ensure that children have access to the digital environment in a way that is non-discriminatory and inclusive and takes into account children’s developing capacities and the particular circumstances of children in vulnerable situations.[4] This should apply also in situations where AI systems are involved. Whereas efforts should be undertaken to respect, protect and fulfil the rights of each and every child in an education setting, targeted measures may be needed to address specific needs, recognising that AI systems have the potential both to increase children’s vulnerability and to empower, protect and support them.[5]

In the context of Article 14 ECHR, positive obligations of States could include measures of ‘reasonable accommodation’ to correct “factual inequalities”.[6] Positive action, or temporary special measures, may involve measures to prevent or compensate for disadvantage suffered by groups exposed to discrimination and intolerance and to facilitate their full participation in all fields of life.[7] Member States should therefore ensure that education institutions use AI systems in a way that is inclusive.[8] States should also make efforts to enhance the use of information and communication technology by closing the gender digital divide and to promote the equality of opportunities and outcomes for all children.[9] In addition, systems such as facial recognition, used as part of a proctoring AI system designed to monitor student behaviour during online exams, can exhibit biases and lead to intersectional discrimination, including on grounds of race and gender.[10] Under the ECHR, any difference in treatment must pursue a legitimate aim and be proportionate.[11] Thus, particular attention should be given to the use of AI systems in selection and exam procedures, in the interest of avoiding discriminatory outcomes.

In addition, limited access to AI systems and tools can prevent individuals or groups from experiencing their benefits and advantages, resulting in disadvantages in various sectors including education. AI literacy, which might be considered an extension or specialisation of digital literacy, should be included in the basic education curriculum from the earliest years, taking into account children’s developing capacities.[12] This includes technical competencies, content creation skills, and critical understanding of online risks and opportunities. Efforts should focus on schools, child-focused organisations, and parents or guardians, ensuring a safe and inclusive digital environment. Digital education policies should not disadvantage children who lack resources at home or live in institutions. Special support should be provided to children with limited or no digital access, including those from socio-economically disadvantaged backgrounds and children with disabilities. States should also work to bridge the digital divide and the gender gap in technology, ensuring equal opportunities for all children, regardless of their background, and with a special focus on girls, in accessing and benefiting from digital tools, including AI systems.[13]

 

[3] 1989 UN Convention on the Rights of the Child, Article 3. On the “best interests of the child”, see Neulinger and Shuruk [GC], No. 41615/07, 6 July 2010, §§ 49-56. The Court also places first “the child’s best interests”, which may override, depending on their nature and seriousness, those of the parents (ibid, §134).

[5] Council of Europe Guidelines on Children’s Data Protection in an Education Setting (2020), Consultative Committee of Convention 108 (T-PD), T-PD(2019)06BISrev5, para 5.4.

[8] Committee of Ministers Recommendation CM/Rec(2019)1 on preventing and combating sexism, in particular II.G. ‘Education institutions’.

[10] GEC/CDADI Study (2023), p. 24.

[11] For example, changes to a university access system that led to differential treatment amounted to a violation of Article 14, in conjunction with Article 2 of Protocol No. 1, despite being intended to rapidly improve the quality of higher education. The unforeseeable application of the new system, coupled with the absence of corrective measures, rendered its implementation disproportionate to that aim – see Altınay v. Turkey, No. 37222/04, 9 July 2013, § 60.