UNESCO’s AI Supervisory Network Turns Toward Practical Cooperation

UNESCO’s Global Network of AI Supervisory Authorities is placing greater emphasis on the practical work of overseeing artificial intelligence across borders. At UNESCO’s fourth Global Forum on the Ethics of AI in Riyadh, held from 14 to 17 September 2026, the network brought together authorities to examine how audit methods, shared experience and institutional cooperation can support AI supervision. Led by the Saudi Data & AI Authority and Luxembourg’s National Commission for Data Protection alongside UNESCO, the network has expanded since the previous Global Forum in Bangkok to include 42 AI supervisory authorities, according to UNESCO.

A central part of the Riyadh discussions was a workshop on AI audit and evaluation. Rather than treating broad risks as labels alone, the session focused on converting them into assessable elements: defining what is to be tested, choosing a suitable method, collecting evidence and communicating findings. This approach matters because an audit is not simply a declaration that a system is acceptable or unacceptable. Participants considered the limits of simple pass-or-fail assessments, which may obscure the seriousness of a finding or fail to show where additional scrutiny is needed. The workshop’s emphasis was on a documented path from a clearly identified risk to findings that can be explained and compared.

A practical case involving an AI-based hiring tool gave participants an opportunity to consider the different parts of a system that may be relevant during an audit. These included the data used, the way the model is coded, the production and checking of outputs, and the assignment of responsibility when a system fails.

International coordination was another major theme. During a panel on 16 September, representatives from Cuba, Luxembourg, Saudi Arabia and Thailand discussed the role of cooperation among supervisors. The discussion reflected a practical challenge identified by the network: AI systems and their supply chains can cross national borders, while incidents may affect more than one jurisdiction at once. In that setting, sharing methodologies, knowledge and experience can help authorities address common supervisory challenges. Participants also highlighted the value of common approaches and trust between supervisory bodies and market actors, while recognising that authorities continue to operate under their own national legal frameworks.

The network’s next phase focuses on cooperation among institutions working in different legal environments. Greater consistency may help reduce unnecessary duplication while supporting cooperation between jurisdictions. Capacity building was particularly relevant for countries still developing AI regulatory and supervisory frameworks, and participants stressed the importance of exchanging both successful experiences and lessons learned. UNESCO said the priorities identified by members include information sharing, international cooperation, capacity building, and greater alignment and harmonisation of approaches.

GNAIS was launched at the third UNESCO Global Forum on the Ethics of AI in Bangkok in 2025 and is part of the implementation of UNESCO’s 2021 Recommendation on the Ethics of Artificial Intelligence, adopted by 193 Member States. Its stated purpose is to connect public supervisory bodies, build institutional capacity and provide access to practical exchanges, regulatory tools and specialised training. The discussions in Riyadh reflected a growing focus on moving from broad discussion toward practical cooperation.