Google found hackers using AI agents to build and run exploits based on software patches
Google Finds Hackers Using AI Agents to Build and Run Exploits From Software Patches
Google's Threat Intelligence Group (GTIG) released a report saying that threat actors' abuse of AI has shifted from manual prompting to deploying open-source autonomous multi-agent attack frameworks such as ARTEX, CAIRN, CyberStrikeAI and PentAGI to carry out multi-stage intrusions at machine speed, and that observed actors include financially motivated groups and suspected China-aligned espionage actors. The report says attackers also use LLMs and agent frameworks to automate reverse engineering of patches through differential analysis, rapidly weaponizing newly disclosed vulnerabilities and compressing the time window from patch disclosure to actual exploitation, with the case being the PaperCut NG/MF vulnerability in September this year.
The report documents the practical trend of attackers moving from manual prompting to multi-agent frameworks and using LLMs to accelerate patch reverse engineering, which defensive teams can use to assess the pressure on the time window from patch disclosure to exploitation.