IGF 2023 WS #469 AI & Child Rights: Implementing UNICEF Policy Guidance

Tuesday, 10th October, 2023 (04:30 UTC) - Tuesday, 10th October, 2023 (06:00 UTC)
WS 4 – Room B-1

Artificial Intelligence (AI) & Emerging Technologies
Chat GPT, Generative AI, and Machine Learning

Organizer 1: Vicky Charisi, University College London, UCL, UK
Organizer 2: Steven VOSLOO, 🔒
Organizer 3: Randy Gomez, HONDA Research Institute, Japan
Organizer 4: Daniella DiPaola, MIT, Media Lab

Speaker 1: Bernhard Sendhoff, Private Sector, WEOG, Global Director, Honda Research Institute
Speaker 2: Joy Nakhayenze, Civil Society, African Group, Tiija Rinta, UCL, Institute of Education, UK
Speaker 3: Judith Okonkwo, Civil Society, African Group
Speaker 4: Ruyuma Yasutake (student, 16 years old) and Tomoko Imai, Civil Society, Asia-Pacific Group
Speaker 5: Dominic Regester, Civil Society, WEOG, Director of the Center for Education Transformation, Salzburg Global Seminar


Vicky Charisi, Technical Community

Online Moderator

Daniella DiPaola, Technical Community, Western European and Others Group (WEOG)


Daniella DiPaola, Technical Community, Western European and Others Group (WEOG)


Round Table - 90 Min

Policy Question(s)

A. How can recommendations and guidelines for AI and Children’s Rights be embedded in industry strategies and translated into concrete technical specifications for AI products that protect children’s privacy, safety, cyber-security, inclusion and non-discrimination and their rights in the digital world?

B. How can we ensure children’s inclusion in the whole cycle of product design, development, implementation, and evaluation?

C. How can governments and intergovernmental institutions support a transparent interaction between policy, industry, and civil society and ensure accountability?

What will participants gain from attending this session? Participants will have the chance to see a use-case scenario with a working AI system that was developed based on UNICEF’s Policy Guidance on AI and Child’s Rights. More specifically, participants will see:

- A tangible example (the Haru robot) of the implementation of the UNICEF policy recommendations on AI and Child's Rights;

- How to apply the lessons learned so far in their own work;

- How we translated ethical principles to technical specifications for an AI system;

- Lessons learned about the methods for including children from different parts of the world, including remote areas with little technological readiness;

- Lessons learned about methods for collaboration between stakeholders (UNICEF - HRI - EC - Academia, Schools and Hospitals);

- Insights from practitioners (educators) and a young participant (student) and the impact of the process on the educational goals;

- Reflections on the process in relation to different AI technologies such as Generative AI and Virtual Reality.


The rapid development of Artificial Intelligence (AI) places children in a constantly changing environment with new applications that utilize Generative AI, Social Robots and Virtual Reality impacting their online and offline lives. To ensure the development of responsible AI for children, UNICEF developed a Policy Guidance on AI and Children’s Rights, drawing on inputs from international experts, governments, industry and children. UNICEF invited organizations to pilot the proposed recommendations and share lessons. Responding to UNICEF’s call, the HONDA Research Institute and the Joint Research Centre of the European Commission has built a multi-stakeholder partnership (academia, industry, schools and intergovernmental institutions) – spanning various areas (Asia, Africa, Europe and the USA) and disciplines (roboticists, psychologists and educators) – to explore techniques for the implementation of UNICEF’s recommendations. This work has been recently featured by the IEEE as one of the tangible examples for responsible AI for children. The roundtable aims to reflect on the experience of the team, including real-world challenges and success stories, to surface concrete guidance to policymakers and developers on how to create child-centred AI systems. The session will provide an overview of UNICEF’s Policy Guidance and will present the current results of its implementation, demonstrating how specific principles are being translated into technical requirements. We have invited an educator from Uganda, a student from Tokyo and a VR developer from Nigeria that co-designs solutions with children to discuss how we consider the perspectives of children from different cultures, socio-economic contexts, and schools with different technology readiness in the process of designing an AI-enabled robot while serving specific educational goals. We contextualize our initiative within the global policy scene, sharing and developing insights applicable to related technologies. We will conclude the session with the participation of the audience and a reflection, including the presentation of our next steps.

Expected Outcomes

After the end of the workshop we aim to publish a report authored and edited by the organizing committee with the description of the session, an initial description of our case study, and a focus on the three Policy Questions about the role of industry on the development of responsible AI, children’s inclusion as a prerequisite for the development of AI for children, and the role of intergovernmental institutions for the transparent interaction among different stakeholders. We also aim to include short inputs written by the invited speakers and to consolidate the notes of the rapporteur and the issues that will be raised during the discussion. We aim to disseminate this report among industry, policy institutions, governments, intergovernmental institutions, academia, and schools.

Hybrid Format: We aim to ensure a balanced experience of online and in-person participation. Organizers and invited speakers will join in a combinatory scheme of online and on-site participation. We plan to schedule the invited talks in turns online and on-site. We will have two moderators, one online and one in person, who will have an internal communication channel for the coordination between the online and on-site coordinator. Out of three speakers, two will be on-site and one online with at least two of the organizers will be on-site. In addition to the facilities and tools provided by the Forum, we aim to conduct short polls among the participants with the use of Slido.com for which both on-site and online participants will participate online. We will have a screen where online participants will be visible to the room of the session. We consider using a telepresence robot.