US lawmakers examine restrictions on Chinese AI models over national security concerns.
Chinese AI models gain US users through lower costs and improving performance.
Congress weighs procurement limits as AI rivalry between the US and China intensifies.
American legislators are weighing steps to limit the growing use of Chinese artificial intelligence models by US companies, as rivalry between Washington and Beijing over AI supremacy intensifies on both sides, CNBC reported.
Chinese models have been winning over US firms by narrowing the performance gap with American counterparts while undercutting them on price. The Trump administration accused Chinese entities in April of running industrial-scale campaigns to steal US AI systems and said it would explore ways to hold foreign actors accountable. Beijing, meanwhile, is separately looking at restricting overseas access to its own leading models.
Two House committees, the Committee on Homeland Security and the Select Committee on China, launched a joint investigation in April into the growing adoption of Chinese-developed AI. Their initial move was to write to companies including Cursor and Airbnb over their exposure to Chinese-built systems. Cursor, which is set to be acquired by Elon Musk's SpaceX for sixty billion dollars, used Chinese model Kimi, developed by Moonshot AI, to build its Composer 2 product. The company declined to comment. Airbnb said its AI activity ran overwhelmingly on US-origin models, adding that a limited number of Chinese open-source models it used operated only through approved US-based service providers.
The limits of a potential ban
Andrew Garbarino, chairman of the Homeland Security committee, told CNBC that Chinese AI was no longer merely catching up but racing to close the gap in capabilities that would shape the future of cybersecurity, describing reports of a Chinese open-weight model matching leading US systems in vulnerability discovery as highly alarming. The State Department said Chinese models were designed to advance Beijing's narratives, censor dissent and reflect Communist Party ideology. China's embassy in London rejected the characterisation as baseless and malicious.
Analysts and officials, however, say curbing adoption will be legally and practically difficult. Since Chinese open-source models are freely available online, an outright ban could run into First Amendment protections around speech, according to Kyle Chan of the Brookings Institution. Daniel Remler of the Center for a New American Security said the administration may also worry that action against Chinese models could harm start-ups that rely on them or dampen broader support for open-source AI. Both analysts suggested procurement rules discouraging companies that do business with the government from using Chinese AI, alongside publishing findings about associated security risks, as more workable approaches. Andy Ogles, chairman of the Subcommittee on Cybersecurity and Infrastructure Protection, warned that if the US failed to act, Chinese models risked becoming the default foundation of the global digital economy, with censorship and stripped-out safety guardrails built in from the ground up.
