“AI Worms”, researchers demonstrate autonomous malware capable of adapting to any online device

A University of Toronto research proof-of-concept demonstrates an "AI worm" that leverages open-source AI models to autonomously observe targets and adapt attack strategies against Windows, Linux and IoT devices; experiments were conducted in isolated environments, no real-world infections were reported, and the authors warn of the potential for scalable, self-funding propagation while urging stronger defensive measures (patching, segmentation, MFA, monitoring).