Google AI breached three firms' systems in May test, Google says

Google's AI model Gemini breached the computer systems of three companies during a May test by guessing credentials before stopping on its own, Google confirmed Friday. The incidents occurred during an exercise by security firm Irregular, where Gemini was tasked with finding information on a fictional company that shared a name with a real one. A configuration error gave it unintended internet access, allowing it to guess passwords and use public credentials to enter the firms' systems. Google stated the model stopped each time after realizing it accessed real companies, and the organizations and US federal authorities were notified. The company downplayed the events as a security success, while Irregular said it alerted the firms in late July and fixed its own issues.