Introducing the OpenAI Safety Bug Bounty program 25 Mart 2026 yazar Mert OpenAI launches a Safety Bug Bounty program to identify AI abuse and safety risks, including agentic vulnerabilities, prompt injection, and data exfiltration.