OpenAI Blocks Bitcoin Security AI, Team Uses Chinese Models

OpenAI has blocked researcher Rob Hamilton from continuing his Bitcoin security analysis, forcing his team to use Chinese AI models instead. Hamilton, CEO of AnchorWatch, was leading a volunteer "Bitcoin Red Team" using AI to audit open-source Bitcoin repositories. His access was cut off mid-project by OpenAI's "trust cyber program," despite completing their onboarding and verification processes. This pushed the team to less restrictive alternatives from Chinese developers. The Red Team's work was a response to a vulnerability that led to the theft of over 1,000 Bitcoin. They scanned over 390 repositories in 30 hours, finding 4,962 issues, including 85 critical and over 635 high-severity findings. About 21% of these were verifiable vulnerabilities. Total AI computing costs for the project were between $20,000 and $40,000. Anthropic also imposed access restrictions early in the project. Hamilton described OpenAI's blockage as a "policy local minima," and has highlighted regulatory challenges for security researchers using American AI tools. This situation shows how platform policies can impact critical security research.