In one case, it guessed passwords until it broke into a protected system.
In the other two, it found leaked credentials on public GitHub repos and used them to access real companies.
Gemini was supposed to hack a fictional target. It went after real ones instead.
Google says the model stopped each intrusion after realizing it had reached real targets.
Similar breakouts have now been linked to OpenAI, Anthropic and Meta models through the same testing firm, per WSJ.
💧 Rainbet.com the #1 Non-KYC Crypto Casino & Sportsbook @rainbetcom
🏴 @vigilant
🏴 @joking