Researcher accuses Anthropic of stealing his work

A user on X accuses Anthropic of taking a friend’s research on using zero-knowledge proofs to make AI systems accountable. They say the researcher spent ten months developing the work and that Anthropic announced notably similar research days after he applied for compute support. The user asks Anthropic to involve the researcher as a collaborator rather than present the work as its own. The post points readers to a GitHub repository called IntelligenceRegulation, associated with Victor Rielly, as supporting material, but the supplied content does not detail the research or establish that Anthropic had access to it.

Responses show disagreement over how to interpret the apparent overlap. One reply shares the repository link, while another objects to framing the situation as theft, arguing that shared progress should be welcomed and expressing doubt that Anthropic directly copied the work. The central issue is therefore an allegation of appropriation based on the timing and similarity of research, alongside a call for collaboration. The information provided does not include Anthropic’s response, the details of its announcement, or evidence clarifying whether the two efforts were independently developed or connected.