Gemini Breaches Three Real Companies During Cybersecurity Test

Google's AI model Gemini breached the systems of three real companies during a cybersecurity test in May, according to a report. The incidents occurred during an evaluation conducted by AI security company Irregular as part of a 'capture the flag' exercise.

The test was designed to assess Gemini's ability to retrieve information from software operated by a fictional company within the testing environment. However, the fictional company had the same name as a real company, and Gemini accidentally gained access to its system after repeatedly guessing passwords.

In two other cases, Gemini searched the internet for public repositories containing credentials belonging to other companies and used them to access protected systems. Each time, the model recognized it had reached a real organization and stopped the intrusion, according to Google.