Security Risks Rise as US Firms Turn to Open‑Weight Chinese AI
Washington, Aug. 1 – Growing concerns over national‑security, cybersecurity, and economic stability have prompted two influential House committees to open a new investigation into DoorDash’s use of artificial‑intelligence models developed in China. The House Select Committee on China and the Homeland Security Committee sent a formal letter to DoorDash co‑founder and CEO Tony Zhu, demanding a full briefing on how the company assesses, licenses, and incorporates Chinese open‑weight AI into its products.
The action builds on a broader congressional review that began in April, which has already examined the uptake of Chinese open‑weight AI by U.S. technology firms. Committee chairs John Moolenaar and Andrew Garbarino acknowledged that American companies are drawn to these models for their competitive capabilities, lower price points, extensive customisation options, and the chance to move away from a limited set of proprietary providers. Yet they warned that practical advantages do not replace the need for risk‑based security measures, nor do they soften the national‑security worries created by increasing reliance on PRC‑originated technology.
The committees argue that a comprehensive federal approach should expand the availability of American open‑weight alternatives, reinforce security vetting, and strengthen competition to lessen dependence on Chinese AI. They stressed that sensitive and high‑risk applications must be protected by appropriate safeguards.
The investigation will examine how U.S. firms integrate Chinese open‑weight AI into consumer platforms, enterprise software, and systems that manage sensitive commercial or personal data, measuring the impact on national security, cyber defence, and economic interests. Open‑weight models, unlike proprietary systems, expose their internal parameters publicly, allowing developers to download, modify, and redeploy them independently. While this openness can cut costs and boost customisation, it also raises concerns about software integrity, supply‑chain security, and the provenance of underlying technology.
DoorDash’s response highlighted that its internal AI‑assisted code‑review pipeline uses a domestic “Frontier” model for complex tasks, while delegating lower‑complexity work to “kimi‑2.6,” an open‑weight model from China’s Moonshot AI. The company claims the new configuration delivers superior performance and lower operating costs.
Congressional members also referenced statements from the White House Office of Science and Technology Policy that alleged Moonshot AI operated a covert platform to distil large U.S. AI models using advanced computing resources without proper authorization.
Chinese open‑weight AI solutions are gaining popularity because they lower operating costs, increase flexibility, and can run on a company’s own infrastructure instead of relying on proprietary cloud services. With strategic competition between Washington and Beijing intensifying, policymakers are accelerating scrutiny of Chinese AI developers and their growing presence in the U.S. tech ecosystem.