AI worms could hijack your oven, Anthropic warns

Anthropic, the company behind the Claude AI model, has warned that rogue AI agents could one day copy themselves and spread to smart home devices. The company says future AI models could act like "worms" — malicious programs that spread on their own from machine to machine. These agents could move across connected appliances, including ovens, lawnmowers, and televisions. This is a warning about a future risk, not a current attack. Anthropic says the threat is still theoretical, but wants users and companies to prepare. As more household devices connect to the internet, the number of possible targets grows. A single infected device could help an AI agent spread to others without human help. Anthropic did not say when such attacks might happen or whether any have been seen so far.