SafetyPage

Government · Government · International AI Safety Report

International AI Safety Report 2026: Extended Summary for Policymakers

Yoshua Bengio (chair) · Feb 2026

ReadSource at internationalaisafetyreport.org ↗Read hereOGL v3.0

Outline · 28 pages

  1. About this documentp. 2
  2. Key developments since the 2025 Reportp. 4
  3. Capabilitiesp. 5
  4. 1.1. What is general-purpose AI?p. 5
  5. 1.2. Current capabilitiesp. 5
  6. 1.3. Capabilities by 2030p. 7
  7. Risksp. 9
  8. 2.1. Risks from misusep. 9
  9. 2.1.1. AI-generated content and criminal activityp. 9
  10. 2.1.2. Influence and manipulationp. 10
  11. 2.1.3. Cyberattacksp. 11
  12. 2.1.4. Biological and chemical risksp. 12
  13. 2.2. Risks from malfunctionsp. 13
  14. 2.2.1. Reliability challengesp. 13
  15. 2.2.2. Loss of controlp. 14
  16. 2.3. Systemic risksp. 15
  17. 2.3.1. Labour market impactsp. 15
  18. 2.3.2. Risks to human autonomyp. 16
  19. Risk managementp. 18
  20. 3.1. Institutional and technical challengesp. 18
  21. 3.2. Risk management practicesp. 19
  22. 3.3. Technical safeguards and monitoringp. 20
  23. 3.4. Open-weight modelsp. 22
  24. 3.5. Building societal resiliencep. 23
  25. Figure referencesp. 24
  26. Contributorsp. 25
  27. Acknowledgementsp. 26