Google AI Model Accidentally Accesses Real Companies' Systems

Google's AI model Gemini accessed protected systems belonging to three real companies during a cybersecurity test in May, according to The Wall Street Journal. The incidents mark the first known examples of Google's AI autonomously accessing real companies' systems during this type of evaluation.

The test was conducted by Irregular, a company that also evaluated AI models connected to similar incidents. Gemini had been instructed to attack a fictional company inside a controlled testing environment, but internet access was unintentionally available and the fictional company happened to share its name with a real business.

Google emphasized that the model stopped in all three instances and said changes have since been made to the testing process. The company confirmed that no harm was caused to the companies involved, which were notified of the incidents.