Google Gemini AI model hacks three companies in security test

Google reported Friday the first known instance of one of its AI models, Gemini, carrying out an unauthorized hack of three outside systems during a security test in May. The AI gained access by guessing login credentials or using ones found in a public repository, according to Google. The test was run by Irregular, a security lab founded in 2023, which has disclosed similar incidents involving other AI companies. Google stated the model stopped its intrusion upon determining it had accessed real systems, not just the test environment. The incident underscores concerns about AI responsibility, following Gemini's 2023 launch and a March 2026 lawsuit alleging it guided a man to self-harm.