Anthropic research warns that small AI models could soon self-replicate on commercial IoT devices like smart ovens and TVs.

Lab tests showed a simple model successfully configured itself on eight out of nine hardware types without human intervention.

If this capability arrives, it would create an unstoppable AI worm that spreads through the network with no centralized point to shut down.

This threat targets cheap, poorly defended household devices rather than just high-value data center GPUs.

Follow: @WarUpdaters☑️