3:01 PM ET · Oct 13

Anthropic warns small AI models may soon self-replicate across smart home devices

Anthropic's Frontier Red Team says small AI models may be months away from self-replicating on insecure IoT devices such as smart ovens and off-brand Android TV boxes, without needing a data center. In controlled tests, models configured inference on unfamiliar hardware and chained a simplified replication path; Anthropic said such a system would be challenging to shut down because there is no single central host.

Source: Anthropic

Also covering this story

On X

Anthropic said its experiments suggested small AI models may be months away from being capable enough to self-replicate on commercial internet-of-things devices, such as smart ovens and off-brand Android TV boxes
— Reuters (@Reuters) Oct 13, 2026

Anthropic warns rogue AI agents may ‘self-replicate’ to hijack everyday devices
— Financial Times (@FT) Oct 13, 2026

A worthwhile (and unnerving) reminder from Anthropic that we should be paying attention to AI ability thresholds, not just steady climbs In this case, the threshold is the capability needed for a self-replicating AI worm on IoT devices like off-brand TVs
— Ethan Mollick (@emollick) Oct 13, 2026

Back to The Wire · Subscribe

Anthropic warns small AI models may soon self-replicate across smart home devices

Also covering this story

On X

Know what matters in AI.