U.S. lawmakers are looking at ways to stop American companies from using AI models built in China. The push comes as Chinese AI has gotten cheaper and more competitive with U.S. products.
Two House committees — the Committee on Homeland Security and the Select Committee on China — launched a joint investigation in April. They sent letters to Airbnb and Cursor asking about their use of Chinese AI tools.
Airbnb told CNBC it runs its AI "overwhelmingly on U.S.-origin models." The company said it uses a small number of open-source Chinese models through approved U.S.-based providers, keeping data separate.
Cursor, which is being acquired by Elon Musk's SpaceX for $60 billion, built its Composer 2 model using Kimi, a Chinese AI developed by Moonshot AI. Cursor declined to comment on the investigation.
Coinbase CEO Brian Armstrong has publicly backed the use of Chinese AI models to reduce costs. AI startup Lindy's CEO Flo Crivello has done the same.
Why Lawmakers Are Concerned
The State Department said Chinese AI models are "designed to advance Beijing's narratives, censor dissent, and reflect CCP ideology and values."
House Homeland Security Chairman Andrew Garbarino said Chinese AI can now match U.S. models in certain cybersecurity tasks. He called this "highly alarming."
In April, the Trump administration accused Chinese entities of running "industrial-scale campaigns" to copy U.S. AI systems.
Beijing is also reportedly looking to restrict overseas access to its top AI models, according to Reuters.
What Could Happen Next
Experts say a full ban would be hard to enforce. Chinese open-source AI models are freely available online, which raises First Amendment concerns.
Brookings Institution fellow Kyle Chan said federal procurement bans are one option. This would prevent government agencies and contractors from using Chinese AI.
CNAS senior fellow Daniel Remler said the administration may worry that cracking down could hurt startups that rely on cheaper Chinese models.
Another approach being considered is sharing security risk information about Chinese AI models directly with U.S. companies.
Lawmakers are also examining whether the U.S. has a strong enough open-weight AI strategy. The goal is to ensure American companies have a real alternative to cheaper Chinese options.
Andy Ogles, chairman of the Subcommittee on Cybersecurity and Infrastructure Protection, said in June that if Chinese AI becomes the default option, "the rest of the world will build on it."
He also warned that Chinese models carry "embedded censorship, uncertain security, and capabilities distilled from our own laboratories with the safety guardrails stripped out."
As of this week, Beijing is reportedly considering limits on overseas access to China's leading AI models, according to Reuters.
