Benchmarks play a central role in the development and governance of models, yet it is often unclear whether they actually measure the concepts they purport to measure (e.g., reasoning, refusal). We ad...

Meet Dr. Sanmi Koyejo: Stanford computer scientist, AI researcher, and AAS plenary speaker working to make artificial intelligence a more trustworthy partner in scientific discovery.

GSK Careers is hiring a AI/ML Research Scientist, Responsible AI in London, United Kingdom. Review all of the job details and apply today!

The Algorithmic Fairness Across Alignment Procedures and Agentic Systems (AFAA) workshop aims to spark discussions on rethinking fairness in AI alignment procedures and agentic system development.

Angelina Wang, Michelle Phan, Daniel E. Ho, Sanmi Koyejo. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

Large language models aren't effective for many languages. Scholars explain what's at stake for the approximately 5 billion people who don't speak English.

In collaboration with The Asia Foundation and the University of Pretoria, this white paper maps the LLM development landscape for low-resource languages, highlighting challenges, trade-offs, and strat...

There is an urgent need for a more robust and comprehensive approach to AI evaluation. There is an increasing imperative to anticipate and understand ...

We have an opportunity to build systems that don’t just replicate our current inequities. Will we take them?

Machine Learning from Human Preferences

Independent experts present the second draft of the General-Purpose AI Code of Practice, based on the feedback received on the first draft, published on 14 November 2024.

We articulate fundamental mismatches between technical methods for machine unlearning in Generative AI, and documented aspirations for broader impact that these methods could have for law and policy. ...

We introduce novel methods for adapting diffusion models under differential privacy (DP) constraints, enabling privacy-preserving style and content transfer without fine-tuning. Traditional approaches...

We introduce novel methods for adapting diffusion models under differential privacy (DP) constraints, enabling privacy-preserving style and content transfer without fine-tuning. Traditional approaches...