Agents
157d agoOpenAI Launches Safety Bug Bounty
By launching a global Safety Bug Bounty, OpenAI invites researchers worldwide to uncover agentic vulnerabilities, prompt injection flaws, and data exfiltration risks in its models.
The Signal
The initiative signals a broader industry shift toward proactive AI safety testing, encouraging collaboration across academia, industry, and governments to strengthen AI systems against misuse and unintended behavior.