Researchers at the University of Toronto have demonstrated an AI worm capable of targeting any online device, raising concerns about cybersecurity vulnerabilities. This development highlights the potential risks associated with AI in the realm of cyber threats.