OpenAI CEO Sam Altman just publicly admitted that AI agents are becoming a problem; says

The role will tackle cybersecurity, biosecurity, and self-improving AI risks , a crucial step as AI's dual nature becomes apparent. OpenAI is ...