On Wednesday, U.S. government officials pulled a Chinese AI search tool that had been briefly deployed on the Federal Register website, Reuters reported. The removal followed a social‑media post that highlighted a contradiction: the National Archives was using one of Alibaba’s Qwen AI models while the FBI had warned that such models illegally copy U.S. frontier models.
Key Takeaways
- The Federal Register website briefly offered a Qwen AI search option before it was removed.
- The National Archives’ use of Alibaba’s model conflicted with the FBI’s stance on “industrial‑scale distillation.”
- Reactions from the Information Technology and Innovation Foundation and legal scholars highlight concerns over policy consistency and security.
| Metric | Value | Why it matters |
|---|---|---|
| AI model used | Alibaba’s Qwen | Contradicts FBI’s warning that such models copy U.S. frontier models. |
| Date of removal | Wednesday (after September 15) | Shows rapid response to public scrutiny. |
| Agency involved | National Archives (Federal Register) | Responsible for public access to federal documents. |
What Happened?
The Federal Register, the official daily publication for U.S. federal regulations, had briefly incorporated a Qwen AI model to let users search public comments on proposed rules. The feature appeared on September 15, as shown in a screenshot shared by X user @tleilax___, and was confirmed by an archived source code snapshot. By Wednesday, the model was removed, leaving the site without the AI‑enhanced search capability for at least a day.
The Contradiction: National Archives vs FBI
While the National Archives was offering the Qwen model, the FBI had earlier this month named Alibaba among six Chinese firms accused of “industrial‑scale distillation.” The agency claims that this practice helps China’s AI industry cut costs and accelerate development, potentially undermining U.S. leadership. The FBI’s stance is that such models illegally replicate U.S. frontier models, a claim that directly clashes with the National Archives’ deployment.
Timeline of the Qwen Deployment
It remains unclear when the National Archives began offering the Qwen search option. The only concrete dates are the screenshot from September 15 and the removal on the following Wednesday. No official statement from the Archives explains the brief window of availability or the decision to pull the tool.
Reactions from Stakeholders
Daniel Castro, president of the Information Technology and Innovation Foundation, called the situation an “insane” disconnect.
“It’s an insane disconnect for a U.S. agency to use an Alibaba model while the FBI encourages stakeholders to use only American models,”he told Reuters. The National Archives, the White House, and the FBI have all declined to comment on the incident.
Security Implications and Expert Views
Georgetown University Law Professor Anupam Chander noted that whether the Qwen deployment posed a security risk depends on its training data and whether any U.S. data was processed by Alibaba‑controlled systems. Senator Marker Warner (D‑Va.) echoed this concern, emphasizing the importance of understanding the data flow into foreign-controlled AI infrastructure.
Why it matters
For technologists and policy makers, the episode underscores a growing tension between rapid AI adoption and national security safeguards. The National Archives’ brief use of a foreign model highlights how public‑facing government platforms can become arenas for contested AI practices. The FBI’s warning about industrial‑scale distillation points to a broader strategy to protect U.S. intellectual property in the AI domain. Together, these events illustrate the need for clear, consistent guidelines on which AI models can be deployed in federal systems.
Frequently Asked Questions
Why did the National Archives choose Alibaba’s Qwen model?
The source does not provide a rationale for the selection. The brief deployment suggests it may have been a pilot or experimental feature.
What does “industrial‑scale distillation” mean in the FBI’s context?
It refers to the process of extracting knowledge from large proprietary models to build new AI systems, a practice the FBI claims helps Chinese firms replicate U.S. frontier models.
Has the FBI issued any guidance on which AI models federal agencies can use?
While the FBI has encouraged stakeholders to use only American models, it has not released a formal policy specifying permissible AI providers for federal use.
Could the Qwen model have accessed U.S. data?
Georgetown Law Professor Anupam Chander said that the security risk hinges on whether U.S. data was processed by Alibaba‑controlled systems, but no evidence of such data flow was reported.
What will happen next for the Federal Register’s search functionality?
There is no official statement. The removal suggests a pause or reevaluation of AI integration on the platform.
Bottom Line
The Federal Register’s brief use of Alibaba’s Qwen AI model, followed by a swift removal, highlights a clash between federal agencies’ eagerness to adopt AI tools and the FBI’s cautionary stance on foreign models. The incident raises questions about policy consistency, data security, and the future of AI deployment in U.S. government systems. As the AI landscape evolves, clear guidelines and transparent decision‑making will be essential to balance innovation with national security.
Reporting based on: original source.
Comments
Post a Comment