Google Warns Fully Automated AI Cyberattacks Have Arrived and Will Worsen Fast

The Decoder · rss · 2026-11-17

Google's Threat Intelligence Group has for the first time observed a cyberattack with no human involvement: an AI framework called exploitagent, found on an exposed server, automatically compared code before and after patches, built exploits, and ran them on its own. Google warns such fully automated attacks will get worse fast.

Original post →

More from AGI Musings

AGI Musings channel →