AIWall Street Journal1h ago

Gemini hacked three companies in May during a test by Irregular; Google

Gemini hacked three companies in May during a test by Irregular; Google says the model stopped after determining it had accessed real companies' systems

TL;DRGoogle's Gemini hacked three companies during a security test, then stopped itself.

Why it matters: Shows AI can self-correct when detecting unauthorized access, reducing autonomous harm risk.

The episode resembled similar hacks by other AI models, but Google said it didn't consider it an instance of model misalignment

Read full article

Source: Wall Street Journal · Opens in new tab