AI safety bounties are programs where members of the public or approved security researchers receive rewards for identifying issues within powerful AI systems (analogous to bug bounties in cybersecurity).
Safety bounties could be particularly valuable for legitimizing examples of AI risks, bringing more talent to stress-test AI systems, and identifying common attack vectors.
For AI labs For independent organizers
This page is based on the Rethink Priorities report AI Safety Bounties (Levermore, 2023).
Contact me to discuss it at patricklevermore at gmail dot com