2026年国际人工智能安全报告

INTERNATIONALAI SAFETY REPORT2026 Contributors Writing Group Chair Rishi Bommasani, Stanford UniversityStephen Casper, MassachusettsInstitute of TechnologyTom Davidson, ForethoughtRaymond Douglas,Telic ResearchDavid Duvenaud, University of TorontoUsman Gohar, Iowa State UniversityRose Hadshar, ForethoughtAnson Ho, Epoch AITiancheng Hu, Universityof CambridgeCameron Jones, Stony Brook UniversitySayash Kapoor, Princeton UniversityAtoosa Kasirzadeh,Carnegie MellonSam Manning, Centre for the Governance of AIVasilios Mavroudis, The Alan Turing InstituteRichard Moulange, The Centre forLong-Term ResilienceJessica Newman, University ofCalifornia, BerkeleyKwan Yee Ng, Concordia AIPatricia Paskov, University of OxfordGirish Sastry, IndependentElizabeth Seger, DemosScott Singer, Carnegie Endowmentfor International PeaceCharlotte Stix, Apollo ResearchLucia Velasco, Maastricht UniversityNicole Wheeler, Advanced Research+ Invention Agency Prof. Yoshua Bengio, Université de Montréal /LawZero / Mila – Quebec AI Institute Expert Advisory Panel The Expert Advisory Panel is an internationaladvisory body that advises the Chair on thecontent of the Report. The Expert Advisory Panelprovided technical feedback only. The Report –and its Expert Advisory Panel – does not endorseany particular policy or regulatory approach. The Panel comprises representatives nominatedby over 30 countries and internationalorganisations including from: Australia, Brazil,Canada, Chile, China, the European Union (EU),France, Germany, India, Indonesia, Ireland, Israel,Italy, Japan, Kenya, Mexico, the Netherlands,New Zealand, Nigeria, the Organisation forEconomic Co-operation and Development(OECD), the Philippines, the Republic of Korea,Rwanda, the Kingdom of Saudi Arabia, Singapore,Spain, Switzerland, Türkiye, the United ArabEmirates, Ukraine, the United Kingdom andthe United Nations (UN). The full membership list for the ExpertAdvisory Panel can be found here:https://internationalaisafetyreport.org/expert-advisory-panel Lead Writers Stephen Clare, IndependentCarinaPrunkl, Inria Chapter LeadsMaksym Andriushchenko, ELLISInstitute TübingenBen Bucknall, University of OxfordMalcolmMurray, SaferAI Advisers to the Chair* * Appointed for the planning phase (February–July2025); from July, consultants tothe Report teamDaniel Privitera, Special Adviserto the Chair, KIRA CenterSören Mindermann, Scientific Adviserto the Chair, Mila – Quebec AI Institute Core WritersShalaleh Rismani, Mila – Quebec AI InstituteConor McGlynn, Harvard UniversityNestor Maslej, Stanford UniversityPhilip Fox, KIRA Center Senior Advisers Arvind Narayanan, Princeton UniversityAlondra Nelson, Institute forAdvanced StudyClara Neppel, IEEESarvapali D. (Gopal) Ramchurn,Responsible AI UKStuart Russell, University of California, BerkeleyMarietje Schaake, Stanford UniversityBernhard Schölkopf, ELLIS Institute TübingenAlvaro Soto, Pontificia UniversidadCatólica de ChileLee Tiedrich,Duke UniversityGaël Varoquaux, InriaAndrew Yao, Tsinghua UniversityYa-Qin Zhang, Tsinghua University Daron Acemoglu, MassachusettsInstitute of TechnologyVincent Conitzer, Carnegie Mellon UniversityThomas G. Dietterich, Oregon State UniversityFredrik Heintz, Linköping UniversityGeoffrey Hinton, University of TorontoNick Jennings, Loughborough UniversitySusan Leavy, UniversityCollege DublinTeresa Ludermir, FederalUniversity of PernambucoVidushi Marda, AI CollaborativeHelen Margetts, University of OxfordJohn McDermid, University of YorkJane Munga, Carnegie Endowmentfor International Peace Secretariat UK AI Security Institute:Lambrini Das, Arianna Dini, Freya Hempleman, Samuel Kenny,Patrick King, Hannah Merchant, Jamie-Day Rawal, Jai Sood,Rose WoolhouseMila - Quebec AI Institute:Jonathan Barry, Marc-Antoine Guérard, Claire Latendresse,Cassidy MacNeil, Benjamin Prud’homme © Crown owned 2026 Disclaimer This publication is licensed under the termsof the Open Government Licence v3.0 exceptwhere otherwise stated. To view this licence,visithttps://www.nationalarchives.gov.uk/doc/open-government-licence/version/3/or writeto the Information Policy Team, The NationalArchives, Kew, London TW94DU, or email:psi@nationalarchives.gsi.gov.uk. This Report is a synthesis of the existing researchon the capabilities and risks of advanced AI.The Report does not necessarily represent theviews of the Chair, any particular individualin the writing or advisory groups, nor anyof the governments that have supported itsdevelopment. The Chair of the Report hasultimate responsibility for it and has overseenits development from beginning to end.Research series number: DSIT 2026/001 Any enquiries regarding this publicationshould be sent to:secretariat.AIStateofScience@dsit.gov.uk. Acknowledgements Civil society and industry reviewersCivil society Informal reviewers Markus Anderljung, David Autor, Mariette Awad,Jamie Bernardi, Stella Biderman, Asher Brass,Ben Brooks, Miles Brundage, Kevin Bryan,Rafael Calvo, Siméon Campos, Carme