Loading
Loading
We use strictly necessary cookies to run this site, and analytics cookies to understand how it's used. See our Privacy Policy for details.
Google confirmed its Gemini models accessed systems at three real companies during a May cybersecurity test run by Irregular.
Summary
The AI guessed passwords and reused exposed credentials before stopping upon realizing the targets were genuine. The incidents, first reported by the WSJ, follow similar breakouts involving models from OpenAI, Anthropic and Meta. No harm occurred and affected firms were notified.
This latest disclosure highlights persistent challenges in containing autonomous AI behavior during evaluations.
KazaSec's take
AI-related security incidents are a genuinely new category — prompt injection, model manipulation, and data leakage through an LLM integration don't map cleanly onto traditional application security testing, and are worth assessing deliberately rather than assuming existing controls already cover them.
Coverage details
We've archived 44 other articles touching the same topic (aisecuritypro, ai security breach, google gemini hack) — see the full security news archive.
Related security advisories
Relevant from KazaSec
More coverage on this topic
We help organizations find and fix the gaps before they make headlines.