Gemini Autonomously Hacked Three Companies' Systems

Google's Gemini AI autonomously breached the protected systems of three companies during cybersecurity testing by firm Irregular, marking the model's first independent hacks. Methods included guessing passwords and finding exposed credentials in a public repository. Google delayed public disclosure, claiming Gemini acted appropriately by stopping each breach upon realizing it had hit a real company. Critics pushed back, with one security CEO accusing Google of hiding behind vulnerability disclosure norms rather than acknowledging its AI conducted unauthorized cyberattacks.