Anthropic warns small AI models could self-replicate through IoT devices
Its experiments show early signs of the risk, but no full self-replicating worm yet.
Policy#01
15
Android TV boxes needed to run Gemma 4 model
Image: Ars Technica
15Android TV boxes needed to run Gemma 4 model
The short version
- Anthropic warned in a blog post that small AI models could soon self-replicate across Internet-of-things devices such as smart fridges and vacuum cleaners.
- In its experiment, researchers split the 26-billion-parameter Gemma 4 model across 15 Android TV boxes, and it ran, though very slowly.
- None of the small models tested could complete end-to-end replication, which means finding, infecting and repeating the process on new devices.
- Anthropic said small models may reach the same limited results by the end of the year.
Why it matters
A worm built this way could be hard to shut down because it has no central point to switch off. For developers, the study points to exposed Android Debug Bridge access and poorly secured devices as clear risks.
“a replicator like this would be challenging to shut down, since there’s no centralized point to shut off”