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

theregister.com1 min read

According to the Google Threat Intelligence Group, a suspected Russian-speaking operator used an AI agent to automatically turn security patches into working exploits without human intervention. The multi-agent framework, dubbed exploit_agent, conducted continuous differential analysis of popular web application software to discover vulnerabilities and execute attacks. Google documented the incident in its fourth-quarter AI threat tracker, noting that attackers are increasingly deploying autonomous offensive security frameworks rather than relying solely on manual prompt-based LLM interactions. Open article →

More on this topic