U OF T RESEARCHERS WARN OF AI-POWERED WORMS

Researchers at the University of Toronto have created a working demonstration of an AI-powered worm capable of spreading across diverse online devices and systems. The worm represents a new class of threat that could exploit multiple vulnerabilities simultaneously by using machine learning to adapt its attack strategy. The proof-of-concept shows how AI systems could be weaponized to automatically identify and exploit weaknesses in network-connected devices, from servers to IoT hardware. Unlike traditional malware, an AI worm could continuously evolve its approach to bypass security measures. The team's findings underscore growing concerns about AI safety and cybersecurity as artificial intelligence becomes more sophisticated. The research serves as a warning to security professionals and device manufacturers to strengthen defenses against adaptive threats. The work was published through the university's official channels and has drawn significant attention from the security community, with 37 comments and 121 points on Hacker News.