US government removes Chinese AI model from Federal Register site after FBI warning

Why this matters

• The use of a Chinese AI model on a US government website undercuts the federal government's own security messaging about Chinese AI technology, particularly after the FBI's public accusations of industrial-scale copying.
• The incident raises broader questions about procurement practices across federal agencies, which may not always align with stated national security policies on AI supply chains.
• The removal signals that public scrutiny can force rapid corrective action in government digital services, but leaves unclear whether other agencies are using similar tools.
• This could accelerate federal moves to audit and restrict AI tools deployed on government platforms, potentially affecting future procurement for public-facing services.

Background

The Federal Register is the official journal of the US government, publishing proposed regulations and public comments. Search tools on such sites have been enhanced with AI to help citizens navigate dense text. Alibaba's Qwen models are a family of open-source AI systems developed in China, which have been praised for their capabilities but have drawn increasing scrutiny from US officials. In September 2026, the FBI named Alibaba among six Chinese firms allegedly engaged in 'industrial-scale distillation' — a process where models are trained on outputs from more advanced systems to replicate their capabilities at lower cost. This came amid broader US efforts to limit Chinese AI advancement, including export controls and legal actions. The National Archives is an independent agency, so its procurement choices may not have been reviewed under the same security umbrella as defense or intelligence agencies.

Key perspectives

• The National Archives: Positioned as an agency focused on public access to government documents; it may have selected the Qwen tool for its search quality or cost-effectiveness, potentially unaware of the FBI's security concerns or considering them irrelevant for public comment search. Its silence leaves room for speculation about its decision-making process.
• The FBI and National Security Officials: Emphasize national security risks, viewing Chinese AI models as vectors for intellectual property theft and potential espionage. For them, any use of such models in US government systems is a clear vulnerability, regardless of the application.
• Civil Liberties and Open-Minded Commenters: Some may argue that open-source models like Qwen are not inherently dangerous and that blanket bans could hinder government innovation. They might also point out that the FBI's accusations have not been proven in court.
• Critics and Social Media Users: See the incident as an embarr, buttressing arguments that federal agencies are not consistently enforcing AI security policies. They are likely to push for stricter compliance.

What to watch

• Watch for any official statement from the National Archives explaining the reason for removal and timeline of the Qwen deployment.
• Monitor whether the FBI or Department of Justice issues formal guidance or legal action against Alibaba that could affect all federal agencies' use of Qwen models.
• Look for announcements of government-wide AI procurement reviews or updated security directives from the Office of Management and Budget regarding AI tools in federal websites.
• Watch whether other federal agencies quietly remove or confirm use of Chinese AI models following public attention, indicating a broader crackdown.