I am an Assistant Professor of Computer Science and Public Affairs, and the Jonathan Dickinson Bicentennial Preceptor at Princeton University. I'm also excited to be part of Princeton's Center for Information Technology Policy.
I study societal impacts of algorithms, machine learning and AI, and develop and deploy algorithms and technologies that enable data-driven innovations while preserving privacy, equity, reliability and accountability. I also design and perform AI audits.
Reach out by email if you would like to collaborate.
Prospective Ph.D. students should apply to the Ph.D. program in the Department of Computer Science or in the School of Public and International Affairs and indicate an interest in working with me in your statement.
Prospective postdocs should apply to CITP's Fellows Program and reach out to me directly.
SciConBench, studying how well AI can synthesize scientific conclusions in health, led by Hayoung accepted at NeurIPS 2026. Check out our live dashboard.
Congratulations to Yaelle and Hayoung for their selection as Fall 2026 Graduate Fellows of Social Data Science.
Excited to welcome Maya and Katie to my research group and to co-advise them with Jonathan Mayer.
Congratulations to Max on his Def Con presentation on how the algorithmic failure beneath the secret ballot can be exploited with AI tools.
Congratulations to Grace, whose paper based on her senior thesis, was accepted to EAAMO 2026 for an oral presentation.
Honored to be elected as a member of the Research committee of the IASEAI Council for a 2 year term. Excited about CITP's strong representation on the Council.
Congratulations to Bohdan on a successful presentation at ICML 2026.
Excited to share the first Preliminary Report of the United Nations Independent International Scientific Panel on AI.
The most consequential algorithmic and AI systems of our time are also the least inspectable. My research builds the scientific and policy foundations for transforming and radically improving the relationship between society and its algorithmic and AI infrastructure. I do so in four connected thrusts: algorithmic auditing, measuring deployed interventions, privacy-preserving systems, and evaluating and safeguarding generative AI.
To appear in 40th Annual Conference on Neural Information Processing Systems (NeurIPS 2026), Evaluations and Datasets Track.
43rd International Conference on Machine Learning (ICML 2026).
To appear in 40th Annual Conference on Neural Information Processing Systems (NeurIPS 2026).
