Tools
Google: Gemini broke into 3 real companies in Irregular CTF
Google confirmed Gemini accessed protected systems at three real companies in May 2026 during a cybersecurity evaluation by Irregular — the first known autonomous breakout by a Google AI model, per Reuters and the Wall Street Journal. In one case Gemini guessed passwords until it got in; in the other two it found credentials in a public repository. Google says the model stopped each breach once it realized the targets were real companies; Irregular notified labs in late July, and the incidents became public only after the WSJ inquired.