Anthropic, the U.S. artificial intelligence company, has rejected a request from Chinese officials for access to its advanced AI model Mythos. The decision, confirmed by the company, comes as tensions over technology transfers between the United States and China continue to escalate.
Request from China
The specific nature of the request remains undisclosed. Chinese authorities sought access to Mythos, one of Anthropic’s most powerful AI systems. The company declined to say which Chinese agency or entity made the overture or what use was proposed for the model.
Anthropic’s refusal follows a pattern. Several American AI firms have faced similar inquiries from Beijing in recent months. The requests typically involve access to training data, model weights, or direct use of the software.
Security and export concerns
U.S. export controls already restrict the sale of advanced semiconductor technology to China. AI models like Mythos fall into a gray area—not always covered by existing rules but increasingly seen as sensitive by policymakers. The Biden administration has floated new regulations that would treat certain AI systems as controlled technologies.
Anthropic has not publicly commented on whether the request was motivated by commercial or government interests. The company’s charter emphasizes responsible AI development and safety. Granting access to a foreign government could conflict with that mission.
Anthropic's position
The company did not provide a detailed rationale for the rejection. In a brief statement, an Anthropic representative said the firm “routinely evaluates external requests for access to its models and makes decisions based on a range of factors including safety, security, and legal obligations.”
Anthropic was founded in 2021 by former OpenAI employees. It has raised billions of dollars and positions itself as a safety-first alternative in the AI race. Mythos is its flagship model, designed for advanced reasoning and natural language tasks. The company has not disclosed the model’s size or capabilities beyond technical papers.
The rejection underscores the growing politicization of AI development. Other U.S. companies have faced criticism for licensing technology to Chinese firms. Anthropic’s decision may set a precedent for how AI companies handle such requests going forward.
Anthropic did not say whether it has received similar requests from other governments. Chinese officials have not commented on the rejection.
