Agents
153d agoOpenAI Launches Safety Bug Bounty
OpenAI’s new Safety Bug Bounty invites researchers worldwide to uncover agentic vulnerabilities, prompt injection, and data exfiltration risks in large language models. By offering rewards, the program encourages rapid identification of weaknesses that could be exploited, strengthening global AI safety standards.
The Signal
Participants will help safeguard future deployments of systems like ChatGPT and other generative models.