AI-powered hacking: Claude targets OpenAI in new cyber era

Researchers used Anthropic's Claude AI to identify and exploit vulnerabilities in OpenAI's internal code environment, marking a shift where artificial intelligence actively participates in cyberattacks rather than merely serving as a target. The authorized bug-hunting activity demonstrates how AI can accelerate vulnerability assessments that previously took weeks, with systems capable of analyzing code, automating testing, and suggesting exploitation methods without human fatigue or error. Cybersecurity is evolving into an AI-versus-AI arms race, where attackers and defenders deploy intelligent systems at machine speed, raising concerns about the average enterprise's ability to defend against such sophisticated automated threats.