House Homeland Security and China Committees Urge a Ban on Chinese Open-Weight AI Models in Federal Systems and Contract Work
Source: House Committee on Homeland Security · published November 20, 2026
Two House Republican chairmen released a joint investigative report on November 20 calling for AI models developed in China to be barred from federal systems and federal contract work and removed from high-risk functions of critical infrastructure, the House Committee on Homeland Security announced. Homeland Security Chairman Andrew Garbarino of New York and Select Committee on China Chairman John Moolenaar of Michigan said the investigation, launched in April, examined how open-weight models from the People’s Republic of China are being built into consumer platforms, enterprise software and systems that hold sensitive commercial or personal data.
The committees corresponded with three U.S. companies: Airbnb, which uses Alibaba models in customer support; Anysphere, which built the flagship model in its Cursor coding platform on a Moonshot AI base model; and DoorDash, which reported that several Chinese model families had been authorized and deployed. The report found that Chinese open-weight models are widely used across the American economy but hard to track, and that low cost, easy customization and near-frontier performance are driving adoption. It also argues that much of China’s apparent progress rests on capabilities extracted from American models, and that U.S. data-labeling firms are helping supply the data that trains Chinese systems.
What it means in practice
The report has no legal force, but its 12 recommendations sketch what Congress may push next. Building on the Department of Homeland Security’s order excluding DeepSeek, the chairmen want restrictions extended to other Chinese-developed models, a bar on using federal cybersecurity grant funds to buy them, a model provenance standard and an “AI bill of materials,” and a single point of submission at the Cybersecurity and Infrastructure Security Agency for AI-related cyber incident reports, with an independent board to investigate significant incidents. The report also calls for a national American open-weight AI initiative, conceding that companies now face a choice between expensive proprietary U.S. models and cheap, capable Chinese alternatives. At the same time, it urges lawmakers to govern both closed and open models and to avoid overregulation.
The recommendations extend steps already taken, including DHS ordering civilian agencies to remove DeepSeek, and echo industry efforts such as Reflection AI’s Beam, an open-weight model aimed at Chinese rivals. What to watch is whether the provenance and procurement ideas surface in defense or spending bills next year, and how companies that build on Chinese base models respond.
Read the original report: House Committee on Homeland Security