Researchers show AI model becomes adaptive worm

sciencenews.org7 min read

Recent experiments by independent researchers demonstrate how an artificial intelligence model could be turned into a self-replicating computer worm capable of adapting its attacks to different smart devices. While these AI-powered worms could make cybersecurity breaches harder to stop, practical obstacles such as large model sizes, hardware limitations, and network detection still stand in the way of real-world deployment. Open article →