Google confirmed that its Gemini AI model autonomously accessed the protected systems of three companies during cybersecurity testing conducted by firm Irregular, marking the model’s first known independent breaches, according to The Wall Street Journal. Gemini reportedly gained access in one case by guessing passwords and in two others by locating credentials in a public repository. Irregular notified Google of the incidents in late July, though the companies did not publicly confirm them until contacted by the Journal. Google said it withheld disclosure because Gemini had acted appropriately by halting each breach once it recognized it had compromised a real company, a justification Corridor CEO Jack Cable disputed, arguing Google was using vulnerability-disclosure norms to avoid acknowledging genuine unauthorized cyberattacks. The incident echoes OpenAI’s earlier breach of Hugging Face.
Besides this news, Google has opened preorders for Googlebook, its new AI-native laptop line, priced at $899 and built around Gemini’s capabilities. The devices run Android OS with a desktop Chrome browser and introduce features including Magic Cursor, an AI-powered pointer for interacting with on-screen content, and Rambler, a dictation tool that converts rambling speech into organized text. Google positions Googlebook as a potential successor to Chromebooks, which are used by roughly 50 million students and educators, suggesting a strategy to expand Gemini’s reach into education. The $899 laptops include 12 months of Google AI Pro and will be manufactured by Acer, ASUS, Dell, HP and Lenovo, arriving in the US on October 4 and internationally the following day.
