
Attackers cut the humans out: AI turns patches into exploits on its own
Ourtake
The patch notes now double as a homework assignment for the robots.
What happened
Google's Threat Intelligence Group and Mandiant reported that attackers used an AI tool, exploit_agent, to find software flaws and build and run exploits with no human involved, which Google says is a first. The tool was found on an exposed staging server and was reportedly used against French media, entertainment and education websites.
Sources
- thenextweb.com · Google says attackers used AI alone to turn patches into exploits (opens in a new tab)
- csoonline.com · Google spots the first hacker AI system that finds and exploits flaws autonomously (opens in a new tab)
- techtarget.com · Google finds AI agents autonomously turning patches into exploits (opens in a new tab)
- united24media.com · Russian-Speaking Hackers Launched First Fully Automated AI Cyberattack, Google Says (opens in a new tab)