SafetyPage

Government · Government · International AI Safety Report

International AI Safety Report 2026

Yoshua Bengio (chair) · Feb 2026

ReadSource at internationalaisafetyreport.org ↗Read hereOGL v3.0

Outline · 220 pages

  1. Contributorsp. 2
  2. Acknowledgementsp. 4
  3. Forewordsp. 6
  4. About this Reportp. 9
  5. Key developments since the 2025 Reportp. 10
  6. Executive summaryp. 11
  7. Introductionp. 14
  8. Background on general-purpose AIp. 16
  9. What is general-purpose AI?p. 17
  10. Current capabilitiesp. 26
  11. Capabilities by 2030p. 32
  12. Risksp. 44
  13. Risks from malicious usep. 45
  14. 2.1.1. AI-generated content and criminal activityp. 45
  15. 2.1.2. Influence and manipulationp. 50
  16. 2.1.3. Cyberattacksp. 57
  17. 2.1.4. Biological and chemical risksp. 64
  18. Risks from malfunctionsp. 71
  19. 2.2.1. Reliability challengesp. 71
  20. 2.2.2. Loss of controlp. 76
  21. Systemic risksp. 84
  22. 2.3.1. Labour market impactsp. 84
  23. 2.3.2. Risks to human autonomyp. 89
  24. Risk managementp. 96
  25. Technical and institutional challengesp. 97
  26. Risk management practicesp. 105
  27. Technical safeguards and monitoringp. 120
  28. Open-weight modelsp. 132
  29. Building societal resiliencep. 138
  30. Conclusionp. 146
  31. Glossaryp. 147
  32. How to cite this reportp. 156