Anthropic moves to close loopholes that allow Chinese access to Claude
Anthropic is tightening controls to prevent Chinese users from accessing its AI chatbot Claude through third-party platforms, closing loopholes that bypassed its regional restrictions. The company updated its terms of service and deployed technical measures to block unauthorized API usage from China, amid growing US-China tensions over AI technology.
Background
- Anthropic is the US-based AI startup behind Claude, a leading large language model (LLM) that competes with OpenAI's ChatGPT and Google's Gemini. It was co-founded by former OpenAI employees and is known for its emphasis on "AI safety."
- The article reports that Anthropic is tightening its controls to prevent users in China from accessing Claude through third-party resellers or VPNs. As a US company subject to export controls on advanced AI technology, Anthropic must block access from China; some Chinese developers and businesses had been paying for Claude indirectly via API resellers who routed traffic through non-Chinese servers.
- This matters because it highlights the intensifying US-China tech war over frontier AI models. Washington fears that advanced American AI could help China's military or give its tech giants a competitive edge. Enforcement is tricky: the models themselves are software that users can attempt to access from anywhere.