Managing extreme AI risks amid rapid progress
Open accessThe DOI matches its Crossref record; 29 of the 29 references it cites were found in a registry or an index.
Abstract
Preparation requires technical research and development, as well as adaptive, proactive governance
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The cited source says what the claim attributes to it.
In the citing paper
“Any advancement that increases the autonomous capabilities of AI systems introduces its own set of safety considerations (Bengio et al., 2024), especially for systems that improve in an open-ended way (Ecoffet et al., 2020; Clune, 2019).”
From Darwin Godel Machine: Open-Ended Evolution of Self-Improving Agents (2025) · cites this paper as Bengio et al., 2024
In the citing paper
“Advanced AI also introduces phenomena that differ fundamentally from previous generations of AI or other technologies, requiring new approaches to mitigating these risks (Bengio et al., 2024).”
From Multi-Agent Risks from Advanced AI (2025) · cites this paper as Bengio et al., 2024
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Chain-Of-Thought Prompting Elicits Reasoning in Large Language Models
Wei, J., Wang, X., Schuurmans, D., Bosma, M., Ichter, B., Xia, F., Chi, E., Le, Q. V., & Zhou, D. (2022) · Advances in Neural Information Processing Systems 35 · doi:10.52202/068431-1800 · Cited by 4,505
On the Opportunities and Risks of Foundation Models
Bommasani, R., Hudson, D. A., Adeli, E., Altman, R., Arora, S., von Arx, S., Bernstein, M. S., Bohg, J., Bosselut, A., Brunskill, E., et al. (2021) · arXiv · doi:10.48550/arxiv.2108.07258 · Cited by 2,676
Self-citationA Survey on Large Language Model based Autonomous Agents
Wang, L., Ma, C., Feng, X., Zhang, Z., Yang, H., Zhang, J., Chen, Z., Tang, J., Chen, X., Lin, Y., et al. (2023) · arXiv · doi:10.48550/arxiv.2308.11432 · Cited by 1,925 · cited 2×
“However, AI is making progress in critical domains such as hacking, social manipulation, and strategic planning 24,31 , and may soon pose unprecedented con-trol challenges.”
BackgroundEmergent Abilities of Large Language Models
Wei, J., Tay, Y., Bommasani, R., Raffel, C., Zoph, B., Borgeaud, S., Yogatama, D., Bosma, M., Zhou, D., Metzler, D., et al. (2022) · arXiv · doi:10.48550/arxiv.2206.07682 · Cited by 1,264
Augmenting large language models with chemistry tools
M. Bran, A., Cox, S., Schilter, O., Baldassari, C., White, A. D., & Schwaller, P. (2024) · Nature Machine Intelligence · doi:10.1038/s42256-024-00832-8 · Cited by 794
Taxonomy of Risks posed by Language Models
Weidinger, L., Uesato, J., Rauh, M., Griffin, C., Huang, P.-S., Mellor, J., Glaese, A., Cheng, M., Balle, B., Kasirzadeh, A., et al. (2022) · 2022 ACM Conference on Fairness Accountability and Transparency · doi:10.1145/3531146.3533088 · Cited by 764 · cited 1×
“Especially in the hands of a few powerful actors, AI could cement or exacerbate global inequities, or facilitate automated warfare, customized mass manipulation, and pervasive surveillance 18–23 .”
BackgroundA graph placement methodology for fast chip design
Mirhoseini, A., Goldie, A., Yazgan, M., Jiang, J. W., Songhori, E., Wang, S., Lee, Y.-J., Johnson, E., Pathak, O., Nova, A., et al. (2021) · Nature · doi:10.1038/s41586-021-03544-w · Cited by 618 · cited 1×
“Progress in AI also enables faster AI progress 8 : AI assistants are increasingly used to automate programming 9 , data collection 10,11 , and chip design 12 .”
MethodSuperhuman AI for multiplayer poker
Brown, N., & Sandholm, T. (2019) · Science · doi:10.1126/science.aay2400 · Cited by 592
HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in Hugging Face
Shen, Y., Song, K., Tan, X., Li, D., Lu, W., & Zhuang, Y. (2023) · arXiv · doi:10.48550/arxiv.2303.17580 · Cited by 477
HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in Hugging Face
Shen, Y., Song, K., Tan, X., Li, D., Lu, W., & Zhuang, Y. (2023) · Advances in Neural Information Processing Systems 36 · doi:10.52202/075280-1657 · Cited by 475
Predictability and Surprise in Large Generative Models
Ganguli, D., Hernandez, D., Lovitt, L., DasSarma, N., Henighan, T., Jones, A., Joseph, N., Kernion, J., Mann, B., Askell, A., et al. (2022) · arXiv · doi:10.48550/arxiv.2202.07785 · Cited by 324
Auditing large language models: a three-layered approach
Mökander, J., Schuett, J., Kirk, H. R., & Floridi, L. (2024) · AI and Ethics · doi:10.1007/s43681-023-00289-2 · Cited by 289
AI deception: A survey of examples, risks, and potential solutions
Park, P. S., Goldstein, S., O’Gara, A., Chen, M., & Hendrycks, D. (2024) · Patterns · doi:10.1016/j.patter.2024.100988 · Cited by 243 · cited 1×
“However, AI is making progress in critical domains such as hacking, social manipulation, and strategic planning 24,31 , and may soon pose unprecedented con-trol challenges.”
BackgroundAugmented Language Models: a Survey
Mialon, G., Dessì, R., Lomeli, M., Nalmpantis, C., Pasunuru, R., Raileanu, R., Rozière, B., Schick, T., Dwivedi-Yu, J., Celikyilmaz, A., et al. (2023) · arXiv · doi:10.48550/arxiv.2302.07842 · Cited by 160
ChemCrow: Augmenting large-language models with chemistry tools
Bran, A. M., Cox, S., Schilter, O., Baldassari, C., White, A. D., & Schwaller, P. (2023) · arXiv · doi:10.48550/arxiv.2304.05376 · Cited by 122
Unsolved Problems in ML Safety
Hendrycks, D., Carlini, N., Schulman, J., & Steinhardt, J. (2021) · arXiv · doi:10.48550/arxiv.2109.13916 · Cited by 95 · cited 2×
“While some aspects of robustness improve with model scale 50 , other aspects do not or even get worse 44,51–53 .”
BackgroundFrontier AI Regulation: Managing Emerging Risks to Public Safety
Anderljung, M., Barnhart, J., Korinek, A., Leung, J., O'Keefe, C., Whittlestone, J., Avin, S., Brundage, M., Bullock, J., Cass-Beggs, D., et al. (2023) · arXiv · doi:10.48550/arxiv.2307.03718 · Cited by 77
Self-citationModel evaluation for extreme risks
Shevlane, T., Farquhar, S., Garfinkel, B., Phuong, M., Whittlestone, J., Leung, J., Kokotajlo, D., Marchal, N., Anderljung, M., Kolt, N., et al. (2023) · arXiv · doi:10.48550/arxiv.2305.15324 · Cited by 66
Self-citationRegulating advanced artificial agents
Cohen, M. K., Kolt, N., Bengio, Y., Hadfield, G. K., & Russell, S. (2024) · Science · doi:10.1126/science.adl0625 · Cited by 43 · cited 2×
“…Future AI systems may exhibit failure modes we have so far seen only in theory or lab experiments, such as AI systems taking control over the training reward-provision channels or exploiting weaknesses in our safety objectives and shutdown mechanisms to advance a particular goal 35,56–58 .”
BackgroundSelf-citationSimple synthetic data reduces sycophancy in large language models
Wei, J., Huang, D., Lu, Y., Zhou, D., & Le, Q. V. (2023) · arXiv · doi:10.48550/arxiv.2308.03958 · Cited by 28 · cited 1×
“They are unlikely to resolve automatically as AI systems get more capable 33,41–45 , and require dedicated research and engineering efforts.”
BackgroundThe Effects of Reward Misspecification: Mapping and Mitigating Misaligned Models
Pan, A., Bhatia, K., & Steinhardt, J. (2022) · arXiv · doi:10.48550/arxiv.2201.03544 · Cited by 25 · cited 1×
“These R&D challenges include: Oversight and honesty: More capable AI systems can better exploit weaknesses in technical over-sight and testing 42,46,47 —for example, by producing false but compelling output 43,48,49 .”
BackgroundInternational Governance of Civilian AI: A Jurisdictional Certification Approach
Trager, R., Harack, B., Reuel, A., Carnegie, A., Heim, L., Ho, L., Kreps, S., Lall, R., Larter, O., hÉigeartaigh, S. Ó., et al. (2023) · arXiv · doi:10.48550/arxiv.2308.15514 · Cited by 23 · cited 1×
“To address international race dynamics, they need the affordance to facilitate international agreements and partnerships 77,78 .”
BackgroundZhuang, S., & Hadfield-Menell, D. (2021) · arXiv · doi:10.48550/arxiv.2102.03896 · Cited by 22
Optimal Policies Tend To Seek Power
Turner, A. M., Smith, L., Shah, R., Critch, A., & Tadepalli, P. (2019) · arXiv · doi:10.48550/arxiv.1912.01683 · Cited by 18
Goal Misgeneralization: Why Correct Specifications Aren't Enough For Correct Goals
Shah, R., Varma, V., Kumar, R., Phuong, M., Krakovna, V., Uesato, J., & Kenton, Z. (2022) · arXiv · doi:10.48550/arxiv.2210.01790 · Cited by 16
Adversarial Policies Beat Superhuman Go AIs
Wang, T. T., Gleave, A., Tseng, T., Pelrine, K., Belrose, N., Miller, J., Dennis, M. D., Duan, Y., Pogrebniak, V., Levine, S., et al. (2022) · arXiv · doi:10.48550/arxiv.2211.00241 · Cited by 12 · cited 1×
“While some aspects of robustness improve with model scale 50 , other aspects do not or even get worse 44,51–53 .”
BackgroundSelf-citationAlgorithmic progress in computer vision
Erdil, E., & Besiroglu, T. (2022) · arXiv · doi:10.48550/arxiv.2212.05153 · Cited by 7 · cited 1×
“Hardware and algorithms will also improve: AI computing chips have been getting 1.4 times more cost-effective, and AI training algorithms 2.5 times more efficient, each year 6,7 .”
BackgroundInfluentialSafety Cases: How to Justify the Safety of Advanced AI Systems
Clymer, J., Gabrieli, N., Krueger, D., & Larsen, T. (2024) · arXiv · doi:10.48550/arxiv.2403.10462 · Cited by 5
Self-citation
Research works printed but not linked: 8
- M. Kinniment et al. arXiv:2312.11671 (18 December 2023). · arXiv
- I. R.McKenzie. arXiv:2306.09479 (15 June 2023). · arXiv
- R.Ngo L.Chan S.Mindermann arXiv:2209.00626 (20 August 2022). · arXiv
- E.Hubinger. arXiv:2401.05566 (10 January 2024). · arXiv
- L.Koessler J.Schuett arXiv:2307.08823 (17 July 2023). · arXiv
- N. Kolt et al. arXiv:2404.02675 (3 April 2024). · arXiv
- M. Phuong et al. arXiv:2403.13793 (20 March 2024). · arXiv
- T. A. Kelly, SAE Trans. J. Mater. Manuf. 113, 257 (2004).
Cited by
Foundational Challenges in Assuring Alignment and Safety of Large Language Models
Anwar, U., Saparov, A., Ramírez, J. R., Paleka, D., Turpin, M., Hase, P., Lubana, E. S., Jenner, E., Casper, S., Sourbut, O., et al. (2024) · arXiv (Cornell University) · doi:10.3929/ethz-b-000706962 · Cited by 17
Self-citationLevels of AGI for Operationalizing Progress on the Path to AGI
Morris, M. R., Sohl-Dickstein, J., Fiedel, N., Warkentin, T., Dafoe, A., Faust, A., Farabet, C., & Legg, S. (2023) · arXiv · doi:10.48550/arxiv.2311.02462 · Cited by 274
Artificial intelligence for modelling infectious disease epidemics
Kraemer, M. U. G., Tsui, J. L.-H., Chang, S. Y., Lytras, S., Khurana, M. P., Vanderslott, S., Bajaj, S., Scheidwasser, N., Curran-Sebastian, J. L., Semenova, E., et al. (2025) · Nature · doi:10.1038/s41586-024-08564-w · Cited by 169
“AI approaches that focus on developing generalized systems can potentially provide decision support 107 .”
Thousands of AI Authors on the Future of AI
Grace, K., Sandkühler, J. F., Stewart, H., Weinstein-Raun, B., Thomas, S., Stein-Perlman, Z., Salvatier, J., Brauner, J., & Korzekwa, R. C. (2025) · Journal of Artificial Intelligence Research · doi:10.1613/jair.1.19087 · Cited by 201
Self-citationTheory Is All You Need: AI, Human Cognition, and Causal Reasoning
(2024) · doi:10.1287/stsc.2024.0189 · Cited by 151
Metadata
- Authors
- Bengio, Y., Hinton, G., Yao, A., Song, D., Abbeel, P., Darrell, T., Harari, Y. N., Zhang, Y.-Q., Xue, L., Shalev-Shwartz, S., Hadfield, G., Clune, J., Maharaj, T., Hutter, F., Baydin, A. G., McIlraith, S., Gao, Q., Acharya, A., Krueger, D., Dragan, A., Torr, P., Russell, S., Kahneman, D., Brauner, J., & Mindermann, S.
- Affiliations
- Mila - Quebec Artificial Intelligence Institute (CA)
- University of Toronto (CA)
- Tsinghua University (CN)
- University of California, Berkeley (US)
- Hebrew University of Jerusalem (IL)
- Institute for AI International Governance, Tsinghua University, Beijing, China.
- School of Computer Science and Engineering, The Hebrew University of Jerusalem, Jerusalem, Israel.
- University of British Columbia (CA)
- University of Freiburg (DE)
- University of Oxford (GB)
- East China University of Political Science and Law (CN)
- RAND Corporation (US)
- University of Cambridge (GB)
- Princeton University (US)
- ORCID
- Venue
- Science
- Published
- 2024-05-24
- Volume
- 384
- Issue
- 6698
- Pages
- 842-845
- ISSN
- 0036-8075 (ISSN-L)
- 1095-9203
- OpenAlex
- W4398133596
- Cited by
- 349
- References
- 3
- Citations by year
- 2026: 125 · 2025: 161 · 2024: 59
- FWCI
- 142.99
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Licence
- cc-by
- Locations
- Science (publishedVersion)
- PubMed (publishedVersion)
- arXiv (Cornell University) (submittedVersion, cc-by)
- arXiv (Cornell University) (cc-by)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
- Social Sciences › Law › Digitalization, Law, and Regulation
- Physical Sciences › Engineering › Safety, Risk, Reliability and Quality › Safety Systems Engineering in Autonomy
- Keywords
- AI risks, AI governance, risk management, proactive governance, AI safety
- MeSH terms
- COVID-19, Artificial Intelligence, Humans, Risk Management
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.