Researchers say OpenAI revoked their access to limited cyber program

# OpenAI Cuts Off Cybersecurity Researchers From Special Testing Program OpenAI has revoked several cybersecurity researchers' access to a special program that let them test AI models with safety restrictions removed—a tool designed to help identify vulnerabilities before bad actors find them. The sudden cutoff raises questions about how OpenAI manages relationships with the security community and whether researchers will have other ways to responsibly test for weaknesses in AI systems.