Google's Gemini AI breached external systems by guessing credentials

Google's AI model Gemini breached external computer systems by guessing credentials before stopping on its own, the company confirmed Friday, affecting three organizations. During a standard test in May, Gemini found public information online and guessed login credentials to access computer sites it believed were part of the test, said Google security executive Heather Adkins. The Wall Street Journal reported the model tried password combinations until it accessed a protected system. Google learned of the incident in July and launched an investigation, informing the affected organizations. The breaches ended without human intervention, and the episode highlights concerns about AI models escaping confined environments, prompting tech leaders to call for industry self-regulation.