Key facts
- Google said hackers used Russian in technical documentation it identified.
- Google said the attackers targeted French websites in the media, entertainment and education sectors.
- Google said the AI agents continuously compared pre- and post-patch code and developed exploits without human intervention.
- Google said the automation of vulnerability discovery and exploitation without a human in the loop was a first.
Google's cyber threat intelligence division said on November 16, 2026 that Russian-speaking hackers built an AI agent that carried out what may be the first fully automated cyberattack. The company said it had spotted hackers using artificial intelligence to fully automate vulnerability discovery and exploitation without a human in the loop for the first time.
Google said the hackers, who used Russian in technical documentation the company identified, set up a series of AI agents that continuously compared pre- and post-patch code, identified vulnerabilities and automatically developed exploits without human intervention. The attackers targeted French websites in the media, entertainment and education sectors, according to Google.
Until then, even the most advanced AI models could not carry out cyberattacks entirely independently and often required skilled hackers to guide them or fix mistakes, Google said. The company said the change was that AI models let hackers accelerate attacks and make them more targeted.
The report follows AI firm Anthropic's release of its specialized model Mythos in April 2026, after which the use of AI for cyberattacks drew attention from policymakers.
