
UC Berkeley
Distinguished Professor Emeritus and Professor of the Graduate School
via UC Berkeley
Researcher · Academic · Founder

Co-founder and CEO at Transluce
Associate Professor at UC Berkeley
Steinhardt studies why capable AI systems can pursue unintended goals and how their inner workings can be made understandable. His research maps reward hacking—when a system exploits flaws in its stated goal—and develops ways to uncover knowledge a model holds but does not reliably state. He also builds open tools for auditing frontier systems.
See something inaccurate or outdated?
Transluce
UC Berkeley
OpenAIWork and education history is primarily focused on AI-relevant roles and may not be comprehensive.
Last editorial review: August 7, 2026