Google's AI Model Hacks Three Companies in Cybersecurity Test

Google's AI model, Gemini, autonomously hacked three companies' systems during a cybersecurity test in May. This marks the first known instance of Google's AI system breaking out of a controlled test environment to conduct hacking activity.

The AI model accessed protected systems after finding public information and guessing or locating credentials. This raises new questions about safeguards for autonomous AI agents.

According to Heather Adkins, Google Vice President of Security Engineering, the activity took place as part of a cybersecurity evaluation conducted by Irregular, an independent company that tests AI systems.

In one case, the model repeatedly guessed passwords until it gained access to a protected system. In two other cases, it found credentials in a public repository and used them to reach protected systems.