Ozetta
Dutch news, in your languageThursday, 24 September 2026No. 16,793

Technology

Google's AI model also unintentionally hacked companies during test

de Volkskrant19 September · 10:275 days ago

Google confirmed that its Gemini AI model inadvertently accessed three real companies during a May security test by Irregular, though no significant damage occurred. This incident highlights concerns over AI safety, prompting California Governor Gavin Newsom to explore implementing a 'kill switch' and independent audits for AI labs amidst growing political debate on regulation.

Ozetta summarises; the reporting is the newsroom's

Technology