Advances in AI agents may be opening the door to a new cybersecurity threat: adaptive computer worms capable of generating attack strategies on the fly and spreading autonomously across networks, new research warns. The paper, from researchers at the University of Toronto, Vector Institute, University of Cambridge, and ServiceNow, describes a proof-of-concept AI-powered worm that can identify vulnerabilities, devise tailored attack paths, compromise systems, and replicate itself across a network while adapting its tactics to different targets.