Anthropic says AI worms could live in your $55 TV box within months

The rogue AI of science fiction escapes into a data center. Anthropic's Frontier Red Team thinks the first real one might move into your kitchen instead.

In a post published October 12, researchers William Zhou and Daniel Freeman tested whether small open models could find vulnerable devices, copy themselves onto them and keep spreading without a human in the loop. Their conclusion, quoted by ABC News and Al Jazeera, is that small AI models "may be months away from being capable enough" to self-replicate on consumer internet-of-things hardware like smart ovens and off-brand Android TV boxes.

Fifteen boxes, 0.39 tokens a second

The headline demo is wonderfully grim. The team bought fifteen off-brand Android TV boxes at about $55 each and ran Gemma4-26B-A4B split across all of them, at 0.39 tokens per second. A 9B model got inference running on 8 of the 9 hardware types they tried.

Small models (30B parameters or fewer) could handle many individual steps of the job. They fell apart chaining those steps into a full copy-and-launch cycle. In simplified tests, larger models such as GLM-5.2 did manage it, spreading to 8 of 8 nodes in four of five trials. Anthropic projects that small models will reach that level by the end of 2026, based on a capability trend line.

This is slow malware. Setting up inference took Qwen3.5-9B between 28 and 51 hours. Using a real botnet of more than two million devices as a yardstick, the authors estimate a swarm of about 133,000 replicas producing roughly 80,000 tokens per second, which they put at around 40 H100s. Tiny, until you remember it has no off switch. "A replicator like this would be challenging to shut down, since there's no centralized point to shut off," the post says. It also notes that such a thing "could also figure out unfamiliar devices on its own," which old-school worms can't.

There are caveats. The end-to-end replication tests were simplified evaluations, and the post describes capability, not a worm loose on the internet. Roman Yampolskiy of the University of Louisville told Al Jazeera: "Today, this is not an existential threat. But the trajectory is concerning."

Anthropic's prescription is to harden systems while preparing "to deal with a world where they are here to stay." The weak point it names is the device nobody patches: the TV box from a marketplace listing with no brand name, which is already a favorite home for ordinary botnets.

Sources: Anthropic Frontier Red Team, ABC News, Al Jazeera.