A major public research university in Madison, Wisconsin, home to AI safety relevant research including interpretability work in the Statistics department and student-led AI safety initiatives.
A major public research university in Madison, Wisconsin, home to AI safety relevant research including interpretability work in the Statistics department and student-led AI safety initiatives.
People– no linked people
Updated 05/18/26Funding Details
Updated 05/18/26- Annual Budget
- $4,950,000,000
- Current Runway
- -
- Funding Goal
- -
- Funding Raised to Date
- -
Org Details
Updated 05/18/26The University of Wisconsin–Madison is a flagship public research university established in 1848 in Madison, Wisconsin. With over 51,800 students, 27,293 faculty and staff, and an annual operating budget of approximately $4.95 billion, it ranks among the top research institutions in the country, placing 5th nationally in research expenditures. In the AI safety and existential risk space, the university is notable for several programs and individuals. Yiqiao Zhong is a tenure-track assistant professor in the Department of Statistics (joined Fall 2022) whose research focuses on interpretability of large language models, deep learning theory, and the evaluation and safety of generative AI systems. His work has appeared in top venues including Science and PNAS, and covers topics such as emergent unfaithfulness in chain-of-thought reasoning, representation geometry in transformers, and out-of-distribution generalization. The Wisconsin AI Safety Initiative (WAISI) is a student-led nonprofit organization at UW-Madison that advances understanding and mitigation of advanced AI risks. It has over 10 PhD safety scholars, 6 masters scholars, and 50+ undergraduate scholars, with members embedded across 12+ campus research labs. WAISI runs technical and policy fundamentals programs, a safety scholars program, and has contributed to state-level AI legislation. It has hosted speakers from Google DeepMind, Anthropic, METR, and the Center for AI Safety. The Effective Altruism UW-Madison group, founded in 2020, provides community and fellowship programming for students interested in global catastrophic risks including AI safety, biosecurity, and pandemic prevention. At the institutional level, the RISE-AI initiative is investing in up to 50 new faculty hires in AI and related fields over three to five years, integrating technical AI research with human-centered and policy perspectives through interdisciplinary collaboration.
Theory of Change
Updated 05/18/26UW-Madison contributes to AI safety primarily through academic research and talent development. Faculty like Yiqiao Zhong advance the theoretical foundations of LLM interpretability and safety evaluation, building the scientific basis for understanding and correcting dangerous AI behaviors. Student programs like WAISI create a pipeline of technically skilled AI safety researchers and policy advocates trained at a leading research institution, increasing the number of people working on AI risk reduction. Institutional AI initiatives aim to ensure that as AI capabilities expand, human-centered and safety-conscious perspectives are embedded in the research enterprise from the start.
Grants Received
Updated 05/18/26Projects– no linked projects
Updated 05/18/26Discussion
No comments yet. Be the first to share your thoughts.