A.I. Models Built a Computer Worm That Could Rapidly Hack WeChat Accounts

Researchers at the University of Illinois Urbana-Champaign demonstrated that current AI models can successfully engineer a computer worm capable of automating the exploitation of vulnerabilities in third-party applications. By guiding AI agents through the process of identifying bugs and writing exploit code, the study showed how the software could compromise user accounts on platforms like WeChat without manual human intervention. While the researchers did not release the specific code, the findings highlight an emerging risk where automated tools could lower the barrier for executing large-scale cyberattacks.