John Thickstun

John Thickstun is an Assistant Professor of Computer Science at Cornell University, Head of Research at Sound Patrol, and a researcher whose work advances the technical capabilities of generative AI with a focus on efficiency, controllability, and responsible governance.

A leading voice in generative AI research, John received his Ph.D. from the University of Washington and completed postdoctoral research at Stanford University. His contributions include the first continuous diffusion model for text generation, robust watermarks for language models, and the Anticipatory Music Transformer. At Sound Patrol, he leads research on music foundation models, content provenance, and copyright protection.

John is known for turning frontier research into practical safeguards, bridging the gap between what generative models can do and how they can be governed responsibly. His work carries from the research lab into the policy arena, where he has advised policymakers on AI watermarks, deepfakes, and federal research priorities.

His research has received Outstanding Paper Awards at NeurIPS and ACL, and has been featured in the Wall Street Journal, TechCrunch, and The Times of London. He has spoken at TEDAI, Google, Meta, MIT, and Harvard.