Cybersecurity Policy Report, House Chairmen Seek Info From DoorDash About Use of Chinese AI, (Jul 31, 2026)
Organizations Mentioned:Airbnb, Inc.
The chairmen of two House committees are seeking information from DoorDash, Inc., about its use of AI (artificial intelligence) models developed by companies in the People’s Republic of China (PRC).
Reps. Andrew Garbarino (R., N.Y.) and John Moolenaar (R., Mich.), the chairmen, respectively, of the House Homeland Security Committee and Select Committee on the Chinese Communist Party, warned DoorDash Chief Executive Officer Tony Xu in a letter sent today that “open-weight” AI models developed in China might cost less than proprietary U.S. models but should be evaluated for security risks.
“PRC-developed open-weight models may offer competitive performance, lower operating costs, and greater control over customization and deployment than proprietary models whose weights remain controlled by the developer and are generally accessed through a hosted service or managed cloud platform,” they noted.
“Those characteristics may mitigate certain data governance and vendor dependency risks, but they do not resolve questions about how a model’s capabilities were developed, whether the weights, code, and dependencies were obtained from trustworthy sources, or how the model will behave when given access to proprietary data, source code, development tools, or other sensitive resources,” they said.
They cited reports that DoorDash intended to use Kimi K2.6, an open-weight model developed by Beijing-based Moonshot AI, and advised Mr. Xu to conduct a “meaningful evaluation” that considers “the model and developer, the provenance and integrity of any deployed artifacts, the security of the surrounding software supply chain, the model’s deployment architecture, permissions, and tool access, the results of security and reliability testing, and the company’s ability to replace the model if new intelligence, legal restrictions, or material changes in the developer’s risk profile make continued use unacceptable.”
Reps. Garbarino and Moolenaar asked Mr. Xu to respond by Aug. 14 with the findings of DoorDash’s evaluation and other information about DoorDash’s use of Chinese AI models. The lawmakers previously posed similar questions to Anysphere, Inc., and Airbnb, Inc., as part of their probe into how U.S. businesses are addressing the risks of Chinese AI (CPR, April 29).
News: FederalLegislation DataSecurity AINews