OpenAI Blocks Bitcoin Security Research, Forces Switch to Chinese AI Models
An AI-powered security audit of open-source Bitcoin repositories was abruptly halted by OpenAI in August 2026. Rob Hamilton, CEO of AnchorWatch, had assembled a volunteer 'Bitcoin Red Team' to run the analysis. The team used cutting-edge models, including Kimi K3 from Moonshot AI and GPT Sol with OpenAI's Cyber Harness.
In just 30 hours, they scanned over 390 repositories and found 4,962 issues, including 85 critical ones. However, their access to these models was revoked by both OpenAI and Anthropic, citing policy restrictions.
HAMILTON attributed the blockage to a 'policy local minima' and publicly disclosed the issue on August 9, 2026. He noted that despite completing onboarding and KYC processes, his access was still restricted. As a result, the team switched to Chinese open-source models, specifically Kimi K3.
The incident highlights regulatory challenges faced by researchers using American AI tools for legitimate security work. The Bitcoin Red Team's findings were privately disclosed to repository maintainers following standard responsible disclosure practices.