Publications

« Home

☑ represents archival peer-reviewed publications; * denotes equal contribution.

Book AI as Normal Technology
Arvind Narayanan, Sayash Kapoor
Princeton University Press (in preparation)
Preprint Births are difficult to predict even with rich survey and full-population register data
Elizaveta Sivak, ..., Sayash Kapoor et al.
Preprint (2026)
Preprint Can AI agents conduct open-ended AI research? Early evidence from two case studies · Project page
Peter Kirgis, Sayash Kapoor et al.
Preprint (2026)
Preprint Bridging Predictions and Interventions: An Integrated Framework for Automated Decision-Systems
Inioluwa Deborah Raji*, ..., Sayash Kapoor et al.
Preprint (2026)
Conference ☑ Holistic Agent Leaderboard: The Missing Infrastructure for AI Agent Evaluation · Project page
Sayash Kapoor* et al.
International Conference on Learning Representations (ICLR 2026)
Conference ☑ The Limits of Inference Scaling Through Resampling
Benedikt Stroebl, Sayash Kapoor, Arvind Narayanan
International Conference on Learning Representations (ICLR 2026)
Conference ☑ Towards a Science of AI Agent Reliability · Dashboard · Blog post
Stephan Rabanser, Sayash Kapoor, Peter Kirgis, Kangheng Liu, Saiteja Utpala, Arvind Narayanan
International Conference on Machine Learning (ICML 2026)
Also presented at: Workshop on Failure Modes of Agentic AI at ICML 2026
Conference ☑ FLARE-AI: Flaw Reporting for AI · OpenReview · ICML poster
Shayne Longpre, ..., Sayash Kapoor et al.
International Conference on Machine Learning (ICML 2026)
Workshop Life After Benchmark Saturation: A Case Study of CORE-Bench · OpenReview
Nitya Nadgir, Sayash Kapoor et al.
Second Workshop on Agents in the Wild: Safety, Security, and Beyond at ICML 2026
Workshop Log analysis is necessary for credible evaluation of AI agents · OpenReview
Peter Kirgis, Sayash Kapoor et al.
Workshop on Failure Modes of Agentic AI at ICML 2026
Preprint Seven simple steps for log analysis in AI systems · Inspect Scout
Magda Dubois, ..., Sayash Kapoor et al.
Preprint (2026)
Workshop Open-World Evaluations for Measuring Frontier AI Capabilities · PDF · CRUX · Blog post · OpenReview
Sayash Kapoor et al.
Second Workshop on Agents in the Wild: Safety, Security, and Beyond at ICML 2026
Report International AI Safety Report 2026
Yoshua Bengio, ..., Sayash Kapoor et al.
International AI Safety Report (2026)
A report on the state of advanced AI capabilities and risks written by 100 AI experts
Journal ☑ The 2025 Foundation Model Transparency Index · Project page
Alexander Wan, ..., Sayash Kapoor et al.
Transactions on Machine Learning Research (TMLR 2026)
Online Essay AI Won't Automatically Make Legal Services Cheaper
Justin Curl, Sayash Kapoor, Arvind Narayanan
Lawfare (2026)
Preprint Bridging Prediction and Intervention Problems in Social Systems
Lydia T. Liu, ..., Sayash Kapoor et al.
Preprint (2025)
Report International AI Safety Report
Yoshua Bengio, ..., Sayash Kapoor et al.
International AI Safety Report (2025)
A report on the state of advanced AI capabilities and risks written by 100 AI experts
Preprint A Different Approach to AI Safety: Proceedings from the Columbia Convening on Openness in Artificial Intelligence and AI Safety
Camille François, ..., Sayash Kapoor et al.
Proceedings from the Columbia Convening on Openness in Artificial Intelligence and AI Safety (2025)
Conference ☑ Position: Build Agent Advocates, Not Platform Agents
Sayash Kapoor*, Noam Kolt*, Seth Lazar*
International Conference on Machine Learning (ICML 2025 Position Paper Track)
Journal ☑ AI Agents That Matter · Blog post
Sayash Kapoor*, Benedikt Stroebl*, Zachary S. Siegel, Nitya Nadgir, Arvind Narayanan
Transactions on Machine Learning Research (TMLR 2025)
Conference ☑ The Leaderboard Illusion
Shivalika Singh, ..., Sayash Kapoor et al.
NeurIPS D&B (2025)
Conference ☑ Establishing Best Practices in Building Rigorous Agentic Benchmarks
Yuxuan Zhu, ..., Sayash Kapoor et al.
NeurIPS D&B (2025)
Journal Why an overreliance on AI-driven modelling is bad for science
Arvind Narayanan, Sayash Kapoor
Nature (2025)
Conference ☑ Position: In-House Evaluation Is Not Enough. Towards Robust Third-Party Evaluation and Flaw Disclosure for General-Purpose AI · arXiv
Shayne Longpre, ..., Sayash Kapoor et al.
International Conference on Machine Learning (ICML 2025 Position Paper Track, Spotlight)
Conference ☑ The Reality of AI and Biorisk
Aidan Peppin, ..., Sayash Kapoor et al.
ACM Conference on Fairness, Accountability, and Transparency (FAccT 2025)
Journal ☑ The 2024 Foundation Model Transparency Index
Rishi Bommasani, ..., Sayash Kapoor et al.
Transactions on Machine Learning Research (TMLR 2025)
Journal ☑ The 2023 Foundation Model Transparency Index
Rishi Bommasani, ..., Sayash Kapoor et al.
Transactions on Machine Learning Research (TMLR 2025 Featured certification)
Journal ☑ CORE-Bench: Fostering the Credibility of Published Research Through a Computational Reproducibility Agent Benchmark
Zachary S. Siegel, Sayash Kapoor, Nitya Nadgir, Benedikt Stroebl, Arvind Narayanan
Transactions on Machine Learning Research (TMLR 2024)
Book ☑ AI Snake Oil: What Artificial Intelligence Can Do, What It Can't, and How to Tell the Difference
Arvind Narayanan, Sayash Kapoor
Princeton University Press (2024)
Named one of Nature’s 10 best books of 2024, Bloomberg’s 49 best books of 2024, and Forbes’s 10 must-read tech books of 2024.
Journal ☑ Considerations for governing open foundation models
Rishi Bommasani, Sayash Kapoor et al.
Science (2024)
Journal ☑ REFORMS: Consensus-based Recommendations for Machine-learning-based Science · Blog post
Sayash Kapoor et al.
Science Advances (2024)
Conference ☑ On the Societal Impact of Open Foundation Models · Blog post
Sayash Kapoor* et al.
International Conference on Machine Learning (ICML 2024 Oral)
Conference ☑ A Safe Harbor for AI Evaluation and Red Teaming · Blog post
Shayne Longpre, Sayash Kapoor et al.
International Conference on Machine Learning (ICML 2024 Oral)
Our open letter to AI companies calling for a safe harbor was signed by over 350 academics, researchers, and civil society members.
Journal ☑ How large language models can reshape collective intelligence
Jason W. Burton, ..., Sayash Kapoor et al.
Nature Human Behaviour (2024)
Journal ☑ The Responsible Foundation Model Development Cheatsheet: A Review of Tools & Resources
Shayne Longpre, ..., Sayash Kapoor et al.
Transactions on Machine Learning Research (TMLR 2024 Survey certification)
Preprint Towards a Framework for Openness in Foundation Models: Proceedings from the Columbia Convening on Openness in Artificial Intelligence
Adrien Basdevant, ..., Sayash Kapoor et al.
Preprint (2024)
Journal ☑ Promises and pitfalls of artificial intelligence for legal applications · Blog post
Sayash Kapoor, Peter Henderson, Arvind Narayanan
Journal of Cross-disciplinary Research in Computational Law (CRCL 2024)
Journal ☑ Against Predictive Optimization: On the Legitimacy of Decision-Making Algorithms that Optimize Predictive Accuracy · Blog post
Angelina Wang*, Sayash Kapoor*, Solon Barocas, Arvind Narayanan
ACM Journal on Responsible Computing (JCR 2024)
Also presented at: Philosophy, AI, and Society (2023); Data (Re)Makes the World (2023), ACM FAccT (2023)
Conference ☑ Foundation Model Transparency Reports · Blog post
Rishi Bommasani, ..., Sayash Kapoor et al.
AAAI/ACM Conference on Artificial Intelligence, Ethics, and Society (AIES 2024)
Journal ☑ Leakage and the reproducibility crisis in ML-based science
Sayash Kapoor, Arvind Narayanan
Patterns (2023)
Policy brief Considerations for Governing Open Foundation Models · Blog post
Rishi Bommasani, Sayash Kapoor et al.
Stanford HAI Issue Brief (2023)
Comment The limitations of machine learning models for predicting scientific replicability
M. J. Crockett, Xuechunzi Bai, Sayash Kapoor, Lisa Messeri, and Arvind Narayanan
Proceedings of the National Academy of Sciences Comment (PNAS 2023)
Online essay How to Prepare for the Deluge of Generative AI on Social Media
Sayash Kapoor, Arvind Narayanan
Knight First Amendment Institute (2023)
Conference ☑ Weaving Privacy and Power: On the Privacy Practices of Labor Organizers in the U.S. Technology Industry
Sayash Kapoor*, Matthew Sun*, Mona Wang*, Klaudia Jaźwińska*, Elizabeth Anne Watkins*
ACM Conference on Computer-Supported Cooperative Work and Social Computing (CSCW 2022)
🏆 Impact Recognition Award
Conference ☑ The worst of both worlds: A comparative analysis of errors in learning from data in psychology and machine learning
Jessica Hullman, Sayash Kapoor, Priyanka Nanayakkara, Andrew Gelman, Arvind Narayanan
ACM Conference on AI, Ethics, and Society (AIES 2022)
Conference ☑ Controlling polarization in personalization: an algorithmic framework
L. Elisa Celis, Sayash Kapoor, Farnood Salehi, and Nisheeth K. Vishnoi
ACM Conference on Fairness, Accountability, and Transparency (FAccT) 2019
🏆 Best Paper Award
Journal ☑ Corruption-tolerant bandit learning
Sayash Kapoor, Kumar Kshitij Patel, and Purushottam Kar
Machine Learning (2019)
Journal ☑ A dashboard for controlling polarization in personalization
L. Elisa Celis, Sayash Kapoor, Farnood Salehi, Vijay Keswani, and Nisheeth K. Vishnoi
AI Communications (2019)
Conference ☑ Balanced news using constrained bandit-based personalization
Sayash Kapoor, Vijay Keswani, Nisheeth K. Vishnoi, and L. Elisa Celis
IJCAI Demos Track (2018)