Rishi Bommasani

I am the Society Lead at the Stanford Center for Research on Foundation Models (CRFM).
I am completing my PhD at Stanford Computer Science, advised by Percy Liang and Dan Jurafsky.
Affiliations: Stanford NLP, Stanford AI, and Stanford HAI. Funding: NSF GRFP.

Prior to Stanford, I began research at Cornell (BA Math, BA CS, MS CS) under Claire Cardie.
I am deeply honored to have learned from the late Professor Arzoo Katiyar, who I profoundly miss.
Cornell CS holds a special place in my heart: the department wrote this about my journey. [Profile] [Profile 2]

I research the societal impact of AI, especially foundation models.

Transparency (establish basic facts). holistic evaluation, supply chain monitoring, ecosystem-level analysis, data evaluation
Concepts (introduce new ideas). homogeneous outcomes, trustworthy evaluation, emergent capabilities, data governance
Progress (improve societal outcomes). standards, norms, policy recommendations, regulatory compliance, proportional regulation, power

My research has been featured in The Atlantic, Axios, Bloomberg, Euractiv, Fast Company, Financial Times, Fortune, The Information, MIT Technology Review, The New York Times, Politico, Quanta, Rappler, Reuters, Tech Policy Press, VentureBeat, The Verge and Vox.

Updates

December 2023 Two papers at NeurIPS on measuring systemic failures across ML APIs and inference time for LM APIs. [Ecosystem-level Analysis] [Efficiency Metrics]
November 2023 Policy brief on how to design tiers for proportional regulation of foundation models. [Policy Brief]
November 2023 Several analyses of the Biden Executive Order on AI. [Full Analysis] [Key Statistics] [Open Models] [Implementation Tracker]
October 2023 Percy, Surya and I hosted Stanford HAI's Fall Conference on New Horizons for Generative AI across Science, Creativity and Society. [Website]
October 2023 We launched the Foundation Model Transparency Index. [Website] [Paper] [Blog]
September 2023 Sayash, Percy, Arvind and I hosted the Princeton-Stanford Workshop on Responsible and Open Foundation Models. [Website]
June 2023 Stanford-Princeton response to the NTIA Request for Comment on AI Accountability. [Policy Brief]
November 2022 We released the Holistic Evaluation of Language Models. [Website] [Paper] [Blog]
August 2021 We released On the Opportunties and Risks of Foundation Models! [Paper] [Explainer Video]
August 2021 We inaugurated the Center for Research on Foundation Models (CRFM) as an interdisciplinary initiative spanning 10+ departments! [Announcement Story]