Anthropic claims that small AI models are close to being able to copy themselves across internet-connected home devices, like smart ovens and Android TV boxes, which could let them skip the need for expensive data center infrastructure.
CNBC reports that the AI lab laid out how models could tap the computing power of consumer devices to run copies of themselves and spread to other connected systems. No fully autonomous self-replicating AI exists yet, the company said, but agents are getting better at completing the individual steps such an attack would require.
Anthropic tested the idea using Gemma4-26B-A4B, a small Google model, spread across a network of 15 cheap off-brand Android TV boxes. These kinds of consumer devices tend to have weak security, the researchers noted, which makes them easy to break into.
One technical problem Anthropic flagged: some models are too large to fit on a single device. The company said these could be “split into pieces, or shards” and run across multiple machines at once.
Stopping an attack like this once it’s running would be hard, Anthropic said, because it would be “challenging to shut down” precisely “since there’s no centralized point to shut off.” A spread-out setup like that is a tougher problem than a traditional, centralized system.
Anthropic pointed out that AI agents are “becoming increasingly capable of completing the individual steps in the replication chain (and passing on their adaptations to future copies of themselves).” It also warned that a real attack probably wouldn’t look like one tidy, single-model operation. “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,” Anthropic stated.
Even Anthropic isn’t sure exactly what this would look like in practice. “We are uncertain exactly what form autonomous AI replication will take,” the post said, pointing to a range of possibilities: a “slow, distributed” creep, or something fast and targeted instead. Other scenarios it raised include “a powerful cyber model that breaches datacenters at a high pace, or a model that scams people for compute credits.”
Looking ahead, Anthropic predicted a “future awash with agents running on all types of hardware, including some that aren’t under human control.” Slower replicators matter too, the company said, mainly as an early warning. “The rise of slow replicators would also be a bellwether for more aggressive, faster ones that can breach more hardened machines or be effective API key thieves.”
AI giants simultaneously release powerful AI models and call for industry regulation. David Sacks has pointed out that when AI titans like Dario Amodei, Sam Altman, and Elon Musk call for a slowdown, they are not doing so out of altruism, but with the motivation to capture the industry and exclude themselves from antitrust laws — a conflict that President Donald Trump recognizes. Breitbart News social media director Wynton Hall has written his New York Times bestseller Code Red: The Left, the Right, China, and the Race to Control AI to serve as the definitive guide on how the MAGA movement can create positions on AI that benefit humanity without handing control of our nation to the leftists of Silicon Valley or allowing the Chinese to take over the world.
Read more at CNBC here.
Lucas Nolan is a reporter for Breitbart News covering issues of AI, free speech, and online censorship.