91Â鶹¹ú²ú¾«Æ·×ÔÅÄ

Authors from The Alan Turing Institute : Dr. Mhairi Aitken, Sabeehah Mahomed, Morgan Briggs, Ayça Atabey 

AI is increasingly present within educational contexts where it is used to personalise learning experiences, monitor and track attainment, or assist teachers in planning and delivering lessons. As children are spending more time online and connecting via social media, AI is shaping how they access and receive information about the world and how they form and maintain friendships. Furthermore, AI systems deployed across the public sector are having significant impacts on the lives of children and their families through informing decisions about access to key services in healthcare, education, housing, or criminal justice (to name a few examples). 

Yet, while AI is increasingly impacting all areas of children’s lives, children are largely missing from decision-making processes regarding the design, development, deployment, or governance of AI. This means that often AI technologies are designed without consideration of the particular needs, vulnerabilities, and interests of younger users. Moreover, policy or regulatory responses which have a key role in safeguarding children are all too often developed without insight into children’s actual experiences or interests. 

Without addressing these vital considerations, AI can have significant detrimental impacts on children’s rights. While all but one country around the world has signed up to the United Nations Convention on the Rights of the Child (UNCRC), in 2020, found that of 20 countries with national AI strategies, most included minimal or no mention of children and their rights¹and governance have a role to play in ensuring that children’s rights are upheld at all times, and that AI systems are designed and used to support the best interest of the child. To achieve this, further research is needed to understand children’s varied and changing relationships with AI. 

Overview of Current Work 

Since 2020, the Children’s Rights & AI team at The Alan Turing Institute have worked on exciting new projects exploring children’s rights and AI. Our team worked with UNICEF to pilot their draft policy guidance on child-centred AI. arch we conducted a series of semi-structured interviews with public sector stakeholders in the UK and found that while many stakeholders aspire to engage children in relation to AI, they typically reported that they do not know how. Building off this work, our research article, , examines current approaches designed to address children’s rights in relation to AI systems.  

To expand to a more global perspective, we conducted further in-depth analysis mapping frameworks at the intersections of data-intensive technologies, children’s rights, and wellbeing. These findings were published in our recent report , which includes two heatmaps illustrating our analysis and assessment of 13 transnational frameworks as they relate to three key themes namely, children’s rights, children’s wellbeing, and child-centred recommendations and policies. On a granular level, we explored specific categories of children’s rights (e.g., non-discrimination, best interests of the child) as well as violations of children’s rights in relation to child exploitation. Additionally, we looked at children’s wellbeing and sociotechnical considerations such as the importance of children’s voice and agency, opportunities and risks, and a special consideration for vulnerable and disadvantaged children. Amongst our , we found that most frameworks primarily focused on how governments and policymakers must uphold children’s rights through regulation and guidelines; however, limited mention was made about addressing government’s responsibility to avoid infringing on children’s rights themselves. There was also a lack of distinction between children’s rights and wellbeing, with the two terms used interchangeably at times. Furthermore, less than half of the reviewed frameworks mentioned or recommended a child rights impact assessment (CRIA). 

is also developing best practice in engaging children in AI to inform future approaches to child-centred AI. In an ongoing 2-year project (commenced in 2021) we are collaborating with to engage primary school children (ages 8 – 12) across Scotland on AI and children’s rights. Throughout this project we have held workshops with around 100 children in four schools, and worked with 12 student investigators, selected from the four schools, through online investigator workshops. This project continues to explore how children feel about AI, how they would like to see AI developed in the future, and . In the current phase of the project, children are engaging directly with developers and policymakers to inform decision-making around AI systems that may impact children’s lives. 

Finally, our team at The Alan Turing Institute has been working in collaboration with the Council of Europe’s Steering Committee for the Rights of the Child (CDENF) to conduct a mapping study to assess the need for legally binding frameworks for AI specifically used by children or for systems that affect children up to the age of 18. This mapping study was motivated by three key challenges identified during the high-level launching Conference of the new Council of Europe (Rome, 7/8 April 2022): 

• The lack of legal frameworks that address children’s rights in the context of AI 
• AI systems are often designed in a way that does not consider children’s rights 
• The scientific evidence about the impact of AI on children’s development is still scattered, so is our understanding. 

The results of this mapping study will be published within the first half of this year. 

Key Findings and Recommendations  

Our report demonstrates that there is a need for a multistakeholder approach to international collaboration and knowledge sharing to ensure that these frameworks are utilised effectively in practice. We hope that our work provides a starting point for future research and serves as a useful resource for policymakers and practitioners to explore ways forward in developing and applying child-centred approaches to AI. We look forward to seeing researchers extend similar mappings of frameworks especially to less developed countries, and we have highlighted some additional key areas for further comparative research including data protection, privacy, addressing manipulation and misinformation, and enhancing enforcement measures. 

With many countries anticipating the release of the updated EU AI Act and the Council of Europe [Framework] Convention on AI, Human Rights, Democracy, and the Rule of Law, it remains critical that children and young people remain at the forefront of these conversations. We welcome the Council of Europe’s inclusion of the rights of persons with disabilities and of children in their draft of the [Framework] Convention (see ).  

While we are supportive of the emerging work and efforts at the law and policy levels recognising children’s rights, we believe more needs to be done. As AI is increasingly becoming more prevalent in children’s daily lives, it is vital that children’s rights and children’s voices are at the heart of decision-making processes around the design, development, deployment, and governance of AI systems that impact their lives. Meaningfully involving children in these processes is crucial to ensure that the benefits of AI are realised while protecting children from risks stemming from AI – to create a digital world in which they can thrive.  

_{The ideas and opinions expressed in this article are those of the author and do not necessarily represent the views of UNESCO. The designations employed and the presentation of material throughout the publication do not imply the expression of any opinion whatsoever on the part of UNESCO concerning the legal status of any country, city or area or of its authorities, or concerning its frontiers or boundaries.}