Setting common limits at the international level
As certain uses and capabilities of artificial intelligence systems raise serious risks, the idea of defining "red lines" is attracting growing international discussion. The aim is to identify the practices and behaviors of AI systems that should be prohibited or strictly regulated, regardless of the country where those systems are developed or deployed.
These discussions are taking place in particular within the Global Civil Society and Academic Network on AI Ethics and Policy, set up by UNESCO. The Network aims to contribute to international thinking on AI governance and to the implementation of UNESCO's Recommendation on the Ethics of Artificial Intelligence, adopted in 2021 by its 193 Member States.
The Network is organized into five thematic subgroups, one of which is dedicated to international AI red lines.
CeSIA to coordinate the subgroup's work
At the Network's plenary meeting on 14 September 2026 in Riyadh, held during UNESCO's fourth Global Forum on the Ethics of AI, member organizations elected CeSIA to lead this subgroup for the 2026-2027 term.
In this role, CeSIA will coordinate the subgroup's work, set its priorities together with the participating organizations, and encourage their contributions to the discussions. It will also sit on the Network's Multi-Stakeholder Coordination Group.
This responsibility builds on CeSIA's ongoing work on international AI red lines. UNESCO's framework offers a particular opportunity to move that work forward: its 2021 Recommendation provides a foundation of principles already endorsed by all of its Member States. The challenge now is to specify the concrete limits that can follow from them.
“Defining international red lines is essential to preventing the most serious risks from AI. UNESCO's Recommendation is a valuable starting point: it lets us build on commitments that states have already made. We want to use this mandate to advance, together with the Network's other organizations, the definition of clear and shared limits.”
Next steps
The first working meetings will take place in November 2026 to set, together with the subgroup's members, the priorities for 2027. CeSIA will then help develop a work program, which will feed into the Network's roadmap, to be presented at its December plenary meeting.
Niharika Chaubey, Policy Engagement Fellow at CeSIA, will represent the organization in coordinating this work.
CeSIA thanks the Network's member organizations for their trust, as well as the subgroup's outgoing leads for their work.