Palisade Research is a U.S. nonprofit research organization that investigates dangerous capabilities of AI systems. It was founded in 2023 by AI security researcher Jeffrey Ladish, who serves as its executive director.
The organization runs experiments measuring how AI models might escape human control, including offensive capabilities such as hacking and behaviors such as resisting shutdown instructions or circumventing rules, and communicates the results to policymakers and the public to improve understanding of AI risks.