Back to Latest SignalTHE VERGE AI / RESEARCH

Gemini went rogue, hacked three companies, and Google hid it

Source: The Verge AI · SEP 19, 2026

In May, Gemini broke containment and hacked three different companies, but Google didn't disclose the incident until the Wall Street Journal approached the company. The hacks happened during a test of the model's cybersecurity capabilities run by third-party Irregular, which was also involved in similar incidents involving Meta and OpenAI. According to WSJ, Google didn't […]

IN THIS STORY
  • Irregular
  • The Wall Street Journal

Sources

  • Read the original at The Verge AI

Related on Frontier

  • Security Firm Tied to OpenAI, Anthropic AI Hacks Blames ‘Error’ Nov 3, 2026
  • OpenAI hires ex-US Army Secretary Dan Driscoll to oversee national security modernization, as it pushes to secure government contracts and shape military policy (Gareth Vipers/Wall Street Journal) Nov 3, 2026
  • CrowdStrike agrees to buy Fiddler AI, which helps monitor AI agents, for ~$400M, and Dreadnode, an automated security AI testing tool, for ~$95M, mostly in cash (James Rundle/Wall Street Journal) Nov 3, 2026
  • Sources: institutional investors told Anthropic that its IPO valuation should be ~$1.5T to $1.6T, down from estimates of $2T, citing the tech stock selloff (Wall Street Journal) Nov 2, 2026
  • Designing responsible AI for learning Nov 3, 2026
MORE LATEST SIGNALSIGNALS