OpenAI Expands Bug Bounty to Cover AI Abuse and 'Safety' Concerns

OpenAI launched a Safety Bug Bounty program hosted on Bugcrowd to encourage reporting of issues that pose meaningful abuse and safety risks (such as agentic risks, account integrity violations, and proprietary information abuse). The announcement outlines scope and eligibility, and clarifies differences from its existing Security Bug Bounty, noting that some integrity-violation issues belong in the Security program while general content-policy bypasses without clear safety impact are out of scope.