International AI Safety Report

📅 2025-01-29
📈 Citations: 0
Influential: 0
📄 PDF
🤖 AI Summary
The global AI safety community lacks a unified evaluation framework and empirically grounded benchmark for assessing advanced AI systems. Method: Commissioned jointly by governments from 30 countries, the United Nations, the OECD, and the European Union, this project engaged over 100 interdisciplinary AI experts to independently develop the first multilateral, internationally authoritative AI safety assessment framework. It integrates the Delphi method, cross-institutional evidence synthesis, and multi-stakeholder consensus mechanisms to systematically integrate technical capabilities of state-of-the-art AI systems, emerging risks—including loss of control and malicious misuse—and governance gaps. Contribution/Results: The project delivers the first comprehensive international AI safety benchmark report, structured along three dimensions—capabilities, risks, and governance. It establishes a globally endorsed risk taxonomy and quantifiable safety capability metrics, providing an evidence-based foundation and actionable reference for national and international AI governance policy development.

Technology Category

Application Category

📝 Abstract
The first International AI Safety Report comprehensively synthesizes the current evidence on the capabilities, risks, and safety of advanced AI systems. The report was mandated by the nations attending the AI Safety Summit in Bletchley, UK. Thirty nations, the UN, the OECD, and the EU each nominated a representative to the report's Expert Advisory Panel. A total of 100 AI experts contributed, representing diverse perspectives and disciplines. Led by the report's Chair, these independent experts collectively had full discretion over the report's content.
Problem

Research questions and friction points this paper is trying to address.

Artificial Intelligence Safety
Advanced AI Capabilities
Potential Hazards
Innovation

Methods, ideas, or system contributions that make the work stand out.

Artificial Intelligence Security
Global Expertise Aggregation
Comprehensive Reporting
🔎 Similar Papers
No similar papers found.
Y
Y. Bengio
Sören Mindermann
Sören Mindermann
University of Oxford, OATML
AI safetydeep learningactive learningcausal inferenceCOVID-19
D
Daniel Privitera
T
T. Besiroglu
Rishi Bommasani
Rishi Bommasani
CS PhD, Stanford University
Societal Impact of AIAI PolicyAI GovernanceFoundation Models
Stephen Casper
Stephen Casper
PhD student, MIT
AI safetyAI responsibilityred-teamingrobustnessauditing
Yejin Choi
Yejin Choi
Stanford University / NVIDIA
Natural Language ProcessingDeep LearningArtificial IntelligenceCommonsense Reasoning
P
Philip Fox
Ben Garfinkel
Ben Garfinkel
Director, Centre for the Governance of AI; Research Fellow, University of Oxford
International RelationsAI GovernanceEthics
D
Danielle Goldfarb
Hoda Heidari
Hoda Heidari
Carnegie Mellon University
Responsible AIAI EthicsAI AccountabilityAlgorithmic FairnessAlgorithmic Economics
Anson Ho
Anson Ho
Epoch AI
AIDeep LearningQuantitative MethodsAI Safety
Sayash Kapoor
Sayash Kapoor
CS PhD, Princeton University
ReproducibilityAI agentsSocietal impacts
L
Leila Khalatbari
Shayne Longpre
Shayne Longpre
MIT, Stanford, Apple
Deep LearningNatural Language Understanding
Sam Manning
Sam Manning
Research Fellow, Centre for the Governance of AI
development economicseconomic impacts of artificial intelligence
Vasilios Mavroudis
Vasilios Mavroudis
Research Scientist, Alan Turing Institute
Machine LearningSystems SecurityArtificial Intelligence
Mantas Mazeika
Mantas Mazeika
Center for AI Safety
ML SafetyAI SafetyMachine EthicsML Reliability
Julian Michael
Julian Michael
Scale AI
AI AlignmentComputational LinguisticsNatural Language ProcessingFormal Semantics
J
Jessica Newman
K
Kwan Yee Ng
C
Chinasa T. Okolo
D
Deborah Raji
Girish Sastry
Girish Sastry
Independent
artificial intelligence policymachine learningartificial intelligence
E
Elizabeth Seger
T
Theodora Skeadas
Tobin South
Tobin South
Massachusetts Institute of Technology
Emma Strubell
Emma Strubell
Assistant Professor, Carnegie Mellon University
Natural Language ProcessingMachine LearningGreen AI
Florian Tramèr
Florian Tramèr
Assistant Professor of Computer Science, ETH Zurich
ML SecurityComputer SecurityCryptographyPrivacy
L
Lucia Velasco
N
Nicole Wheeler
D
D. Acemoglu
O
O. Adekanmbi
D
David Dalrymple
T
Thomas G. Dietterich
E
Edward Felten
Pascale Fung
Pascale Fung
Dept. of Electronic & Computer Engineering, the Hong Kong University of Science & Technology
artificial intelligenceconversational AIspeech recognitionnatural language processingAI
P
P. Gourinchas
Fredrik Heintz
Fredrik Heintz
Professor of Computer Science, Linköping University
Artificial intelligenceTrustworthy AIautonomous systemsmulti agent systemscomputational thinking
Geoffrey Hinton
Geoffrey Hinton
Emeritus Prof. Computer Science, University of Toronto
machine learningpsychologyartificial intelligencecognitive sciencecomputer science
Nick Jennings
Nick Jennings
Vice-Chancellor and President, Loughborough University
AIArtificial IntelligenceMulti-Agent SystemsIntelligent Agentsmultiagent systems
A
Andreas Krause
Susan Leavy
Susan Leavy
University College Dublin, Insight Centre for Data Analytics
AI EthicsArtificial IntelligenceNatural Lanugage ProcessingAlgorithmic BiasDigital Humanities
Percy Liang
Percy Liang
Associate Professor of Computer Science, Stanford University
machine learningnatural language processing
T
T. Ludermir
V
Vidushi Marda
Helen Margetts
Helen Margetts
Professor of Society and the Internet, University of Oxford
Political SciencePublic PolicyCollective ActionDigital GovernmentPublic Administration
J
J. McDermid
J
J. Munga
A
Arvind Narayanan
A
Alondra Nelson
C
Clara Neppel
Alice Oh
Alice Oh
KAIST Computer Science
machine learningNLPcomputational social science
G
Gopal Ramchurn
Stuart Russell
Stuart Russell
Macquarie University
Critical criminologylegal theoryhuman rightsrefugeesBrazil
M
Marietje Schaake
B
B. Scholkopf
Dawn Song
Dawn Song
Professor of Computer Science, UC Berkeley
Computer Security and Privacy
Alvaro Soto
Alvaro Soto
Professor Universidad Catolica de Chile
Machine learningcomputer visionrobotics
L
Lee Tiedrich
G
G. Varoquaux
A
Andrew Yao
Y
Ya-Qin Zhang
F
Fahad Albalawi
M
Marwan Alserkal
O
Olubunmi Ajala
G
Guillaume Avrin
C
Christian Busch
A
Andr'e Carlos Ponce de Leon Ferreira de Carvalho
Bronwyn Fox
Bronwyn Fox
A
Amandeep Singh Gill
A
Ahmet Hatip
J
Juha Heikkila
G
G. Jolly
Ziv Katzir
Ziv Katzir
Ben-Gurion University, Department of Software and Information Systems Engineering
Adversarial Machine Learning
Hiroaki Kitano
Hiroaki Kitano
A
Antonio Kruger
C
Chris Johnson
S
Saif M. Khan
Kyoung Mu Lee
Kyoung Mu Lee
Professor, Department of Electrical and Computer Engineering, Seoul National University
Computer VisionMachine LearningArtificial Intelligence
D
D. Ligot
O
Oleksii Molchanovskyi
A
Andrea Monti
N
Nusu Mwamanzi
M
Mona Nemer
N
Nuria Oliver
J
Jos'e Ram'on L'opez Portillo
Balaraman Ravindran
Balaraman Ravindran
Professor of Data Science and AI, Wadhwani School of Data Science and AI, IIT Madras
Reinforcement LearningData MiningNetwork AnalysisResponsible AI
R
Raquel Pezoa Rivera
H
Hammam Riza
C
Crystal Rugege
C
C. Seoighe
J
Jerry Sheehan
H
Haroon Sheikh
D
Denise Wong
Y
Yi Zeng