‘These … clowns’: Why cyber researchers bristle at Anthropic’s ‘self-replicating code’ claims

Anthropic dropped a worrying headline last week, both on X and in a blog on the artificial intelligence company’s website: “AI models may be capable of self-replicating without access to a datacenter.” The post detailed how the company had used the computing power in 15 smart TVs to run an AI model. It also warned that other internet-connected devices could be used to run large language models, riling up cybersecurity experts. “As a cybersecurity expert, the threats they've claimed are not based in fact,” hacker Anthony Jones told Straight Arrow. For folks who don’t know much about self-replicating code, or “computer worms” (read: most normal people), the claim might sound scary. But cybersecurity experts and professionals who make “knowing about worms” their business found the announcement pretty annoying.

Also reported by Straight Arrow.

First reported by Straight Arrow News. This page is a short excerpt; the full text is theirs.