hn.today

Google: AI agent turned a security patch into a exploit with no human help

theregister.com12 points3 comments
Screenshot of Google: AI agent turned a security patch into a exploit with no human help

Google Threat Intelligence Group says a suspected Russian-speaking operator used an autonomous AI framework to turn software security patches into working exploits and run them without human intervention - a first of this kind observed by the group. Investigators found the framework, named exploit_agent, on an exposed staging server. It combined multiple AI models with Russian-language tools and continuously compared popular web applications’ code before and after updates to identify vulnerabilities, build exploits, and execute them. The server also held data stolen from victims and scripts for internal reconnaissance, password brute-forcing, installing webshells, and removing rival attackers’ webshells. Telemetry indicated a focus on French media, entertainment, and education organizations.

The case illustrates a wider shift from attackers asking AI systems for help manually to deploying multi-agent tools for whole stages of an intrusion. Google also observed offensive frameworks including ARTEX, CAIRN, CyberStrikeAI, HexStrike AI, and PentAGI; one PentAGI operation targeted the aerospace sector. Users include financially motivated criminals and suspected espionage groups, including actors with possible ties to China. The tools make it easier to incorporate new exploits and automate attacks at scale, lowering the expertise and effort required. Google expects their use to grow and warns that organizations worldwide may be targeted as these systems become more capable and accessible.

Read on theregister.com3 comments on Hacker News

Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.

More in Security

The daily digest

Today's best Hacker News stories, summarized and screenshotted, one email a day.