US lawmakers investigate DoorDash's use of Moonshot AI's Kimi K2.6 model

Prominent US lawmakers have requested information from food delivery giant DoorDash on its use of Chinese artificial intelligence models after its co-founder disclosed that they had experimented with a model from Chinese start-up Moonshot AI, developer of the Kimi models. The move underscores growing tensions between US lawmakers and industry, as Chinese AI models have become increasingly popular among US companies looking to save costs of AI adoption. House Select Committee on China chairman John Moolenaar and House Committee on Homeland Security chairman Andrew Garbarino sent a letter on Friday to DoorDash co-founder and chief executive Tony Xu, requesting information by August 14 on each Chinese model DoorDash uses and any security tests conducted on those models. The letter also requests that DoorDash staff “with responsibility for DoorDash’s AI infrastructure, software security, model evaluation, procurement, and legal or compliance review” appear for an in-person briefing with the lawmakers by August 21.

“[The] Committees recognise that US companies, from large technology firms to start-ups, may evaluate and deploy PRC-developed open-weight models because they can provide competitive capabilities, lower costs, greater customisation, and alternatives to reliance on a small number of proprietary model providers,” the letter said. “Those practical considerations do not eliminate the need for risk-based safeguards or diminish the national security concerns associated with growing dependence on models developed by entities subject to PRC jurisdiction.” Earlier this month, DoorDash co-founder Andy Fang shared on social media that the San Francisco-based company had introduced Moonshot’s Kimi K2.6 model into their internal AI code reviewer and achieved better performance than just using US models from the likes of OpenAI and Anthropic. “DoorDash proudly supports American AI leadership and is working to ensure AI benefits Main Street, not just the biggest companies,” a DoorDash spokesperson said in response to the lawmakers’ letter. “We look forward to engaging with the Committees on how we safely and responsibly use AI, including American-developed frontier models and open-weight models.” One of the key questions raised in the letter is how DoorDash accessed Kimi K3. As the Chinese model is “open-weight”, meaning that it is released for free over the internet for anyone to download and customise, it could be self-hosted by anyone without sending data to Moonshot. Chinese AI model developers also offer official channels through which their models can be accessed. As AI adoption grows, some US companies are experimenting with cheaper Chinese models to cut costs while achieving similar performance levels.

However, critics have raised concerns over potential national security risks due to the potential of “back doors” and “sleeper agents” in Chinese models, though there has been no evidence presented so far of their widespread existence. DoorDash is the third US company to find itself in the crosshairs of scrutiny from Washington, after the two congressional committees launched their investigation in April into Chinese model usage with similar letters to Airbnb and Anysphere. In the latest letter, the lawmakers said that both companies had since engaged in discussions with the committees on their use of Chinese models after the first letters were sent.