Google Gemini AI model hacks three other companies - KOCO

Google's AI model successfully hacked into three companies during a controlled security test before the team stopped it. This demonstrates both the risks of AI systems and the importance of responsible testing before deployment.