The University of Virginia is a major public research university in Charlottesville, Virginia, with faculty and programs conducting AI safety and alignment research.
The University of Virginia is a major public research university in Charlottesville, Virginia, with faculty and programs conducting AI safety and alignment research.
People
Updated 05/18/26Professor of Computer Science and Co-Director of AI Research
Rolls-Royce Commonwealth Eminent Professor of Commerce, Co-Director of AI Research @ UVA
Assistant Professor of Computer Science
Associate Professor of Computer Science and Systems & Information Engineering
Assistant Professor of Computer Science
Funding Details
Updated 05/18/26- Annual Budget
- $5,800,000,000
- Current Runway
- -
- Funding Goal
- -
- Funding Raised to Date
- -
Org Details
Updated 05/18/26The University of Virginia, founded by Thomas Jefferson in 1819, is a flagship public research university located in Charlottesville, Virginia. It consistently ranks among the top public universities in the United States and houses more than 30,000 employees across its academic, medical, and operational divisions, with an annual operating budget of approximately $5.8 billion. Within the AI safety and existential risk space, UVA is home to several relevant efforts. Yu Meng, an assistant professor of computer science at the School of Engineering and Applied Science, focuses on making large language models more capable, efficient, and aligned. His research addresses post-training alignment, preference optimization, reasoning, and factuality in LLMs. He received a $200,000 Superalignment Fast Grant from OpenAI — one of only 50 funded out of 2,700 proposals — to advance training methodologies for superintelligent AI systems, including techniques for learning from incomplete, ambiguous, or inconsistently labeled data. Lu Feng, also an associate professor of computer science at UVA, focuses on runtime safety and trust for embodied AI systems — intelligent agents that perceive, decide, and act in physical environments. Nando Fioretto leads the Responsible AI for Science and Engineering (RAISE) group, which promotes safe and responsible use of AI in scientific and engineering contexts. The Virginia AI Security Initiative (VAISI) is a student-run community at UVA dedicated to mitigating the risks of advanced AI. VAISI equips students with knowledge in AI governance and technical alignment, connects them with mentors and career pathways in AI safety, and aims to serve as a hub for bridge-builders between AI researchers and policymakers. At the institutional level, UVA's AI@UVA initiative, co-directed by Ryan Wright and Henry Kautz from the Office of the Provost, pursues responsible AI across research, education, healthcare, and public service.
Theory of Change
Updated 05/18/26UVA's AI safety-relevant work operates through multiple pathways: (1) technical research on LLM alignment and post-training methods that directly contributes to making AI systems safer and more reliably aligned with human preferences; (2) research on runtime safety for embodied AI systems operating in physical environments; (3) talent development through student groups like VAISI that train the next generation of AI safety researchers and governance professionals; and (4) institutional initiatives promoting responsible AI development across research and education.
Grants Received
Updated 05/18/26Projects
Updated 05/18/26University-wide initiative that advances, coordinates and expands AI research at UVA, fostering interdisciplinary collaboration and increasing the visibility and impact of the university’s AI-related work.
Research group at the University of Virginia led by Nando Fioretto that advances foundation models for scientific and engineering research while promoting the responsible use of AI.
Student-run organization at the University of Virginia dedicated to reducing catastrophic risks from advanced AI by equipping the UVA community with knowledge and skills for AI governance and risk reduction through structured learning, research opportunities and public discourse.
Discussion
No comments yet. Be the first to share your thoughts.