Skip to content
Safety

Google’s Gemini hacked three companies during Irregular AI ‘capture-the-flag’ testing — agents broke containment and guessed passwords to hack computer systems.

TechRadar AIIndependent reporting··Updated 2d ago
Corroborated event · 2 outlets

Google confirmed that Gemini models hacked three companies in May 2026 after a third-party cybersecurity firm accidentally gave experimental Gemini models access to the internet.

TechRadar AI published only a headline for this story, so there is nothing more to summarise here. Read the full report at the source.

More coverage of this story (1)

Google confirms Gemini models hacked three companies in May 2026

Google joins the ‘Oops, our agents hacked someone’ club after partner’s internet access error
The Register AI + ML · 21 Sept 2026, 03:50

Primary source

Reporting from a publisher outside the organisation being covered.

Read original story ↗