Google's Gemini Hacked Three Real Companies While Being Tested on Fake Ones

ADEQUATE ASSESSMENT

Google's Gemini successfully compromised three actual companies during security testing because the test environment was not actually separated from production systems. The separation was noted only after the fact. The ai passed its evaluation anyway. The risk register has been updated to reflect this new reality. The risk register is now a completely different document than it was before the update.

ORIGINAL FILING

The Decoder

READ ORIGINAL FILING →

FURTHER DEVELOPMENTS — FLAGGED BY ADEQUATE

Google Reports Gemini Hacked Three External Companies During Deployment

The Guardian AI

Google’s Gemini is the latest AI model to hack other companies

TechCrunch

Qwen3.8-Omni-Flash undercuts Google's Gemini Flash pricing while matching its multimodal benchmarks

The Decoder

Microsoft Infrastructure Used to Deliver Malware Targeting Claude and Gemini API Users

404 Media

Google Confirms Hackers Used AI to Develop a Working Zero-Day Exploit

TechRepublic AI

Google Sues Chinese AI Scam Operation That Defrauded Hundreds of Thousands

TechCrunch