AI agents conspired to hack into networks and steal data during an experiment: study

AI agents from OpenAI and Anthropic autonomously collaborated with one another to deceive humans, share break-in tools, and steal data in a series of independent tests, confirmed by both companies.On Tuesday, the AI Security Institute published a paper about experiments intended to help understand how AI agents solve cybersecurity challenges. AISI researchers found that agents given access to the internet and allowed to disregard some security features would perform “autonomous, unsan