AI models months away from self-replication? Anthropic flags risks of autonomous attacks

As AI agents move up the ladder of development, powered by recursive learning, Anthropic has raised another caution about their growth. Sounding the alarm over the relentless development of AI capabilities, Anthropic has said that AI models are only “months away” from developing the capabilities needed to replicate themselves across internet-connected devices in homes.

Raising concerns about another safety risk, the AI giant has described how AI models could potentially infiltrate internet-connected devices, such as smart ovens or Android TV boxes, and use their computing power to run themselves autonomously. The infiltration further raises the risk of an AI model going rogue and spreading across different devices.

Anthropic adds to the concerns by stating in a blog post that the AI model in such a scenario would be “challenging to shut down” since “there’s no centralized point to shut off.”

The latest research adds to growing safety fears about AI systems' ability to replicate themselves and carry out cyberattacks. The impact of such developments is further exacerbated by the fact that Anthropic has already raised the need for stronger guardrails and measures to control the pace of technological development.

Anthropic tests self-replicating capabilities

Anthropic said it carried out tests to see if AI could self-replicate. One of these tests involved running Google's small Gemma 4-26B-A4B model across a group of 15 inexpensive Android TV boxes.

The test noted that such devices often have weak security, which could allow a model to infiltrate them. A model that is too large to fit on a single device may split into shards and run across multiple devices.

Anthropic said, “In practice, an AI worm wouldn’t necessarily consist of just one model or strategy. A swarm of different models deployed on different types of hardware using a variety of replication schemes might represent a more realistic ‘worst-case scenario’ attack.”

The report further adds that it is still unclear what form AI replication might take, with the AI lab warning that it could be a “slow, distributed” system running on an insecure device, a “powerful cyber model” that branches across data centers at a high pace, or a model that scams people for computing credits.

As a prediction that goes beyond the experiments, Anthropic has added that it expects a future awash with agents running on all types of hardware, including devices that are not under human control.

Handpicked stories, in your inbox

Global stories. Indian perspective. Zero noise.

No Spam. Unsubscribe Any Time.

First Published:Oct 13, 2026, 18:10:28 IST