International AI Safety Report (interim 2024; 2025; Key Update Oct 2025; 2026 edition)
Independent expert panel chaired by Yoshua Bengio, mandated by the Bletchley summit; secretariat hosted by the UK AISI · 2025-01-29 (first full report); 2026-02-03 (second edition)
An IPCC-style scientific assessment of AI capabilities and risks, produced annually by an international expert panel. It has become the reference document for summits, though the US stopped endorsing it in 2026.
What it calls for
- Safety evaluations
- International governance
- Transparency
- Alignment research
Scope
~100 experts nominated by 30+ countries, the EU, OECD and UN
What actually happened
Second edition published 2026-02-03 ahead of the New Delhi summit (https://internationalaisafetyreport.org/publication/international-ai-safety-report-2026; arXiv https://arxiv.org/abs/2602.21012). The US withheld its support from the 2026 edition (per TIME, cited at https://en.wikipedia.org/wiki/International_AI_Safety_Report). Bengio was also appointed co-chair of the UN's Independent International Scientific Panel on AI in Feb 2026 (https://en.wikipedia.org/wiki/Independent_International_Scientific_Panel_on_AI). The report is advisory and has no policy mandate.