{"id":399654,"date":"2026-09-23T07:31:55","date_gmt":"2026-09-23T07:31:55","guid":{"rendered":"https:\/\/bizscoreai.com\/blog\/federal-register-pulls-alibaba-qwen-search-tool\/"},"modified":"2026-09-23T07:31:56","modified_gmt":"2026-09-23T07:31:56","slug":"federal-register-pulls-alibaba-qwen-search-tool","status":"publish","type":"post","link":"https:\/\/bizscoreai.com\/blog\/federal-register-pulls-alibaba-qwen-search-tool\/","title":{"rendered":"Federal Register pulls Alibaba Qwen search tool after FBI warning"},"content":{"rendered":"

A small open-weight model from Alibaba’s Qwen family was quietly powering an AI search box on the National Archives’ Federal Register website for at least a day before US officials pulled it. The same week, the FBI had named Alibaba among six Chinese firms conducting what it called industrial-scale distillation of American frontier models.<\/p>\n

What happened on the Federal Register website<\/h2>\n

Visitors to FederalRegister.gov saw an option to search public comments on proposed regulations using a Qwen AI model. A widely shared screenshot posted on X on September 15 by the user tleilax___ documented the option. An archived copy of the site’s source code confirmed that the Qwen model was removed on Wednesday, so the tool was available for at least a day. The National Archives, the independent agency that runs the Federal Register, has not commented on the removal or responded to media requests. The White House and the FBI have also not commented.<\/p>\n

The contradiction with FBI messaging<\/h2>\n

Earlier in September, the Federal Bureau of Investigation named Alibaba among six leading Chinese firms allegedly conducting industrial-scale distillation, which the agency says is helping America’s biggest competitor cut costs and speed up development. Against that backdrop, a federal website quietly using a Qwen model drew sharp criticism. Daniel Castro, president of the Information Technology and Innovation Foundation, called it an insane disconnect for a US agency to use an Alibaba model while the FBI encourages stakeholders to use only American models.<\/p>\n

Which Qwen model was used<\/h2>\n

A report on the Chinese news site Sina.com identified the specific model as a Qwen3-class open-weight release at roughly 0.6B parameters, not one of Alibaba’s flagship models. The deployment was open source, served only to retrieve documents, did not offer complex reasoning, and did not send government data to Alibaba or any third party. Because the model was open weight, it could be downloaded and run locally, giving the hosting agency direct control over queries and outputs.<\/p>\n

Did the deployment pose a security risk<\/h2>\n

Experts quoted in the coverage said the risk was likely low. Georgetown University Law Professor Anupam Chander noted that the Federal Register’s content is already public, so the model was not working with sensitive government information. Senator Mark Warner (D-Va.) added that the question turns on whether any US data was processed by Alibaba-controlled systems, which the open-weight local deployment is designed to avoid. Security specialists told Reuters that running models locally rather than sending queries to external services gives agencies more control over data, which is part of why small open-weight models have become a default AI search tool for many organizations, alongside lower and more predictable pricing. Sina.com’s reporting went further, suggesting that once an American institution takes control of an open-weight model, it is not clear it should still be treated as a Chinese service.<\/p>\n

Where US policy currently stands<\/h2>\n

US lawmakers have not settled whether AI models that originate in China but are controlled by American institutions should face the same restrictions as other Chinese services considered higher security risks. Representative John Moolenaar (R-Mich.), who chairs the House China Committee, has taken the hardest line, stating that no federal government entity should use a Chinese AI model because doing so only makes the federal government more dependent on Chinese AI models, which is not in the national interest.<\/p>\n

Why small open models are at the center of the debate<\/h2>\n

Policy research submitted to the US-China Economic and Security Review Commission in March argued that the US may already be at a structural disadvantage. US export controls are calibrated to constrain frontier training by restricting access to advanced semiconductors. They do not address the small-model deployment cycle, which needs less compute, draws on openly available base models, and produces advantage through application rather than pre-training. If the models that matter most for industrial AI are small, specialized, and open, the current US policy framework could be targeting the wrong layer of the competition.<\/p>\n

Companies including Nvidia and Meta have said they intend to help the US catch up in open model development. Researchers have warned that losing a broad user base to China could set back America’s ability to set the technical standards and norms that govern AI development for years to come. China, which has dismissed accusations of frontier model copying, has urged the US to listen to its own business community or risk putting the US at a further economic disadvantage.<\/p>\n

For US agencies weighing their next move, the Federal Register episode makes the open-weight local deployment argument more concrete, since SEOScanPro’s AI Agent Readiness checks look at exactly this question of whether a deployed model can be inspected, controlled, and run without phoning home.<\/p>\n

FAQ<\/h2>\n

What AI model was used on the Federal Register website?<\/h3>\n

The site used a small open-weight Qwen3-class model at roughly 0.6B parameters from Alibaba, according to a Sina.com report. It was deployed to retrieve documents and did not send government data to Alibaba or any third party.<\/p>\n

Why was the Qwen model removed from FederalRegister.gov?<\/h3>\n

US officials removed the Qwen search option on Wednesday after social media users flagged the apparent contradiction of a US federal site using an Alibaba model at the same time the FBI had named Alibaba among firms conducting industrial-scale distillation.<\/p>\n

Did the Qwen deployment pose a national security risk?<\/h3>\n

Experts quoted in the coverage said the risk was likely low because the Federal Register’s content is already public and the open-weight model was run locally rather than through Alibaba-controlled systems. Lawmakers remain divided over whether any federal use of Chinese-origin models is acceptable regardless of deployment method.<\/p>\n

Related coverage<\/h2>\n