A major public research university whose AI safety-relevant work is centered on the AI+Human Objectives Initiative (AHOI) and Scott Aaronson's computational-complexity-meets-alignment research group, both supported by Open Philanthropy.
A major public research university whose AI safety-relevant work is centered on the AI+Human Objectives Initiative (AHOI) and Scott Aaronson's computational-complexity-meets-alignment research group, both supported by Open Philanthropy.
People
Updated 05/18/26Research Associate Professor of Computer Science
Schlumberger Centennial Chair of Computer Science; Director, Quantum Information Center
Funding Details
Updated 05/18/26- Annual Budget
- -
- Current Runway
- -
- Funding Goal
- -
- Funding Raised to Date
- -
Org Details
Updated 05/18/26The University of Texas at Austin (UT Austin) was founded in 1883 and is the flagship institution of the University of Texas System. As of fall 2025 it enrolls approximately 55,000 students and employs over 4,800 faculty members. In FY2025 it spent $1.37 billion on externally funded research across all disciplines. The most directly AI-safety-relevant program is the AI+Human Objectives Initiative (AHOI), which is building a community of researchers at UT Austin focused on ensuring that the social impact of advanced artificial intelligence is aligned with human goals and values. AHOI draws 13 faculty members from Computer Science (Brad Knox, Peter Stone, Scott Aaronson), Philosophy (Harvey Lederman, Christian Tarsney, Josh Dever, Mark Budolfson), Linguistics (Kyle Mahowald, Jessy Li), and several other departments. AHOI is funded by Open Philanthropy and has also received a grant from Coefficient Giving. Brad Knox, a co-lead of AHOI, is notable for his doctoral research that pioneered the reinforcement learning from human feedback (RLHF) approach now central to large language model training. A second AI safety research thread is led by Scott Aaronson, who holds the Schlumberger Centennial Chair in Computer Science and directs the Quantum Information Center. After a two-year leave (2022-2024) working at OpenAI on the theoretical foundations of AI safety, Aaronson returned to UT Austin to establish a research group studying topics at the intersection of AI safety and computational complexity theory, including interpretability of neural networks, cryptographic backdoors, and out-of-distribution generalization. Open Philanthropy awarded $1,650,000 over three years to support this group. UT Austin also hosts Good Systems, a cross-campus grand challenge initiative that has operated for approximately six years and focuses on developing ethical human-AI systems. Good Systems received a $1 million investment from MITRE and involves faculty and students from more than two dozen UT schools and units. In March 2026 UT Austin hosted the inaugural Texas Symposium on Machine Learning, Responsible AI, and Robotics, which drew over 600 participants from academia, industry, government, and nonprofits. The Institute for Foundations of Machine Learning (IFML), an NSF AI Institute headquartered at UT Austin, conducts foundational ML research to improve accuracy and reliability of AI models, with over $20 million in NSF funding and collaborations with the University of Washington, Stanford, UCLA, and Microsoft Research.
Theory of Change
Updated 05/18/26UT Austin's AI safety programs pursue impact through two complementary paths. AHOI bets that a rigorous, interdisciplinary academic community spanning CS, philosophy, linguistics, and social science can produce foundational research on alignment and safety that informs both technical AI development and policy. Aaronson's group bets that theoretical computer science tools - complexity theory, cryptography, formal methods - can provide rigorous foundations for alignment problems such as interpretability and robustness that currently lack them. Together, these groups aim to build the research base and train the next generation of researchers who will reduce risks from advanced AI systems.
Grants Received
Updated 05/18/26Projects
Updated 05/18/26Interdisciplinary UT Austin initiative building a community of researchers focused on ensuring that the social impact of advanced AI is aligned with human goals and values.
University-wide grand challenge at UT Austin focused on designing, evaluating, and building ethical AI technologies and human–AI systems.
NSF AI Institute headquartered at UT Austin that develops foundational machine learning tools and algorithms to power the next decade of AI innovation and improve the reliability of AI systems.
Discussion
No comments yet. Be the first to share your thoughts.