US Government Removes Chinese AI Model From Federal Register

US government officials removed a Chinese artificial intelligence search tool from the Federal Register website on Wednesday, following public scrutiny. The tool, an AI model from Alibaba's Qwen series, had been briefly deployed on the site, which is managed by the National Archives. This removal occurred after social media users highlighted a perceived contradiction: the National Archives was utilizing an Alibaba AI model while US law enforcement agencies, including the FBI, had previously flagged similar Chinese AI models for potentially engaging in "industrial-scale distillation." The FBI's assessment, made earlier in the month, suggested that such practices by leading Chinese firms could help America's primary global competitor reduce costs and accelerate development timelines in the competitive landscape of artificial intelligence.

It remains unclear precisely when the National Archives, the independent agency responsible for enhancing public access to federal government documents and operating the Federal Register website, first integrated the Qwen AI model as an option for users searching public comments on proposed regulations. The agency has not yet issued a statement regarding the removal of the AI tool, nor did it respond to a request for comment from Ars Technica. Similarly, the White House and the FBI have also refrained from commenting on the matter. The deployment of the Qwen model raised concerns due to the FBI's prior warnings about Chinese AI companies potentially infringing on intellectual property and accelerating their development by leveraging US-developed models. The FBI's designation of Alibaba among six leading Chinese firms involved in "industrial-scale distillation" underscores the geopolitical and intellectual property concerns surrounding the rapid advancement of AI technologies globally.

The Federal Register is a crucial repository for proposed and final administrative regulations published by US federal government agencies. Its website serves as a public portal for citizens and organizations to access and review these important documents. The integration of AI search capabilities was intended to improve user experience and facilitate easier navigation of the extensive public comment archives. However, the choice of an AI model from a Chinese company, particularly one that had been flagged by US intelligence, introduced significant security and trust concerns. The incident highlights the complex challenges faced by government agencies in adopting new technologies while navigating national security interests and international technological competition. The reliance on foreign-developed AI models, especially those with potential ties to state-sponsored research or development, can create vulnerabilities and raise questions about data privacy and intellectual property protection. The swift removal of the Qwen model suggests a cautious approach by the US government to mitigate these risks and ensure the integrity of its public information systems.