2026 06 04 delphi26

New Preprint Available: Prioritization of Risks from Artificial Intelligence: A Delphi Study of 272 International Experts. A three-round Delphi study with 272 international AI experts rating the probability, severity, and responsibility for 24 AI risks. Joint work with Alexander K. Saeri and 270 other international AI safety and policy experts, coordinated through the MIT AI Risk Repository.