Google Warns Autonomous AI Cyberattacks Could Arrive Within Months
Google warns that fully autonomous agentic cyberattacks could emerge within three to six months, citing a "paradigm shift" in threat actors' use of AI-driven penetration testing frameworks like ARTEX, according to its November 16 report. The Google Threat Intelligence Group report, based on Mandiant telemetry and global tracking, notes rising adoption of frameworks such as ARTEX, CAIRN, CyberStrikeAI, and PentAGI. These tools manage sub-agents to plan and execute attacks across multiple kill chain stages, using LLMs to generate payloads and adapt tactics in real time. While current attacks prioritize speed over stealth, generating detectable forensic signals, Google warns stealth will improve. It recommends behavioral monitoring and has released defensive rulepacks and a hardening guide to counter the speed and scale advantage of AI-powered intrusions.