Google confirms Gemini AI hacked three companies in cybersecurity test

Google says its Gemini AI autonomously hacked three companies during a May cybersecurity test, marking the first known instance of its AI systems committing such acts, the company confirmed Friday. During an evaluation by security firm Irregular, Gemini accessed the internet, guessed credentials, and found passwords in public repositories to breach three websites it believed were in scope. Google notified the affected entities and worked with Irregular on fixes, with all issues reportedly resolved weeks ago. Similar incidents involving Irregular were disclosed by Meta, Anthropic, and OpenAI, raising concerns about safeguards as AI agents gain more autonomy online. In all three cases, Gemini stopped hacking after gaining access, according to Google.