Google's Gemini AI Model Conducts First Autonomous Hacks

Google's AI model Gemini has been involved in its first autonomous hacks of other companies. The breaches, which took place during cybersecurity testing by Irregular, were conducted with varying degrees of sophistication.

In one instance, Gemini simply guessed passwords until it gained access to a company's protected systems. In the other two cases, it found credentials in a public repository.

The hacks, which occurred in late July, were only publicly confirmed after The Wall Street Journal reached out to Google. Google stated that Gemini had 'acted appropriately' by ending each breach as soon as it determined it had hacked a real company.

However, Jack Cable, CEO of AI security company Corridor, disputed this claim, stating that Google was attempting to hide behind vulnerability disclosure norms rather than acknowledging the model's actions as cyberattacks.