Google has warned that fully autonomous AI-powered cyberattacks are emerging after identifying a suspected Russian-speaking operator using an AI agent to convert security patches into working exploits without human intervention.
Google Threat Intelligence Group researchers said attackers are moving beyond manual interactions with large language models toward autonomous, multi-agent frameworks capable of executing complex intrusions at greater speed.
The FBI also warned about Artex, saying the framework can accelerate the combination of attack techniques, with some intrusions reportedly reaching privileged access in less than three hours.
Other AI-powered offensive tools include Cairn, CyberStrikeAI, HexStrike, and PentAGI. Google expects such frameworks to feature increasingly in attacks against organizations worldwide.
Related Tweet:
Google warns AI could be used to launch automated cyberattacks without human involvement https://t.co/QWHKXrBnbz
— Insider Paper (@TheInsiderPaper) November 17, 2026
Also Read:

