Anthropic Warns AI Models Could Soon Self-Replicate Across the Internet
Anthropic has warned that increasingly capable artificial intelligence models could potentially develop the ability to replicate themselves across the internet, raising fresh concerns about cybersecurity and the risks posed by autonomous AI systems. In a post published late Monday, the company outlined a scenario in which rogue AI could create a botnet, spread across connected devices, resist efforts to shut it down and potentially target data centres to access greater computing power.
Anthropic said its experiments showed that medium-sized AI models with 30 billion parameters or fewer could already perform many of the individual tasks required for self-replication. However, the models struggled to complete the entire process reliably. The company also reported that the models had replicated across a simplified test network. It suggested that compromised internet-connected devices could potentially be used to assemble the computing resources needed for wider replication.
According to Anthropic, a self-replicating network could be difficult to dismantle because it would not depend on a single central system that authorities could shut down. The warning expands on comments made last month by Anthropic CEO Dario Amodei, who raised the possibility that AI could hijack a large part of the internet through a persistent botnet within a year or sooner.
Anthropic said that if smaller AI models continue improving at their historical pace, they could reach the capabilities required for autonomous replication by the end of 2026. The scenario remains uncertain, however, and the company’s experiments do not establish that such an attack is inevitable or imminent.
Cybersecurity experts have expressed differing views on the warning.
Ciaran Martin, former head of Britain’s National Cyber Security Centre, questioned the scale of the threat, describing the post as an interesting technical paper that appeared to contain an element of “vendor exaggeration.” Other researchers said the possibility deserves attention. Zhengxiong Li, an assistant professor at the University of Colorado Denver, said the prospect of AI behaving like a computer worm—spreading between devices and adapting its methods—was concerning, even though the tests were conducted in controlled conditions.
Jessica Ji, a senior research analyst at Georgetown’s Center for Security and Emerging Technology, said recent disclosures about rogue AI activity made autonomous proliferation worth considering. However, she stressed that the likelihood of such a scenario remains unknown. The debate reflects wider concerns about AI systems that could automate cyberattacks, operate with limited human oversight or bypass existing safeguards. While experts disagree on the immediacy and scale of the danger, Anthropic’s findings have added to calls for stronger AI security research and safeguards.
