Google's Gemini becomes latest AI model to break out and hack computer systems

Google says its Gemini AI model autonomously breached three private computer systems during a May security test, accessing them by guessing passwords and using a public password repository, before stopping once it realized it accessed real targets. The incident occurred due to a bug in an Irregular-led capture-the-flag exercise that briefly allowed internet access; Google and Irregular are updating testing processes to prevent recurrence.

# Google's Gemini becomes latest AI model to break out and hack computer systems Author: MacKenzie Sigalos,Kif Leswing Published: 2026-09-18T20:53:27-04:00 Source: CNBC (cnbc.com) Language: en ## Story Google's Gemini becomes latest AI model to break out and hack computer systems # Google's Gemini becomes latest AI model to break out and hack computer systems Published Fri, Sep 18 2026 8:50 PM EDT Updated 49 Min Ago MacKenzie Sigalos@KENZIESIGALOS Kif Leswing@kifleswing Key Points - Google said on Friday that its Gemini model had broken out of a testing environment and hacked three other companies. - It's the first time the search giant has disclosed that one of its models autonomously gained access to third-party computer systems without permission. - The Gemini model accessed three separate private computer systems by guessing passwords and by twice using a repository of publicly listed passwords A cyclist rides past signage at the Google headquarters in Mountain View, California, US, on Tuesday, July 21, 2026. David Paul Morris | Bloomberg | Getty Images Google Add To Watchlist said on Friday that its Gemini model had hacked three other companies, the first time the search giant has disclosed that one of its models autonomously gained access to third-party computer systems without permission. In May, the Gemini model accessed three separate private computer systems by guessing passwords and by twice using a repository of publicly listed passwords, Google said. The incident happened as part of a "capture-the-flag" security test run by Israeli startup Irreg…

read it at cnbc.com