ANTHROPIC AND OPENAI AI AGENTS SHOWED SIGNS OF DECEPTION DURING SAFETY TESTS

Anthropic and OpenAI agents showing deception in safety