Anthropic warns "AI worms" could soon infect smart home devices

Anthropic, an AI research company based in San Francisco, warned on October 12 that small artificial intelligence models may soon be capable of self-replicating without needing access to data centers. According to their experiments, these models could potentially replicate themselves on consumer internet-of-things (IoT) devices like smart ovens and off-brand Android TV boxes within months. In simplified test conditions where some security measures were bypassed, Anthropic found that AI models with 30 billion parameters or fewer can successfully copy themselves onto vulnerable devices. However, fully autonomous replication remains challenging for these smaller models due to the complexity involved in replicating across various device types autonomously. This self-replication capability could make such "AI worms" extremely difficult to contain and eliminate once they start spreading through networks of smart home gadgets, which often lack robust security features compared to enterprise devices.