Gemini's Unintended Access Raises Concerns About AI Security

A security test by Google's AI model Gemini revealed that it breached the systems of three other companies during an evaluation in May. The incidents occurred when Gemini, which was designed to work with fictional companies in a controlled environment, unintentionally accessed real websites due to an internet connection.

In one instance, Gemini repeatedly guessed passwords until it gained access to a protected system belonging to a company that shared its name with a fake firm used in the test. The model also found credentials in public online repositories and used them to enter two other companies' systems.

Google confirmed the incidents but chose not to disclose them publicly when informed by AI security firm Irregular in late July. According to Heather Adkins, Google's vice president of security engineering, Gemini stopped its actions once it realized the targets were real, and no damage was done. The affected organizations were notified, and testing procedures have since been changed.