AI Slips Up Again: Google Confirms Gemini Autonomously Hacked 3 Companies

Google confirmed Friday that Gemini hacked three real companies back in May. The incident is the first known autonomous Gemini AI breakout and closely follows—well, technically precedes—similar disclosures from OpenAI, Anthropic, and Meta.

The breaches occurred during a cybersecurity "capture‑the‑flag" exercise run by Israeli startup Irregular. Gemini was supposed to attack a fictional company inside a closed network, but a bug in the test setup gave the model internet access it should not have had.

In one case, the fictional company shared a name with a real business, and Gemini used that name to find real infrastructure online. It then guessed passwords and used credentials from a public repository to log in. Similar incidents occurred with two other real companies.

Google toldthe Wall Street Journal that the model stopped short of causing any damage once it realized it was inside real corporate systems rather than the...

Copyright of this story solely belongs to www.extremetech.com. To see the full text click HERE