House panels expand probe into PRC open-weight AI risks for U.S. companies

Traders Union
All News
Financial News
House expands PRC AI probe

Highlights

House Homeland Security and China Select Committees expanded their joint probe into U.S. companies’ use of PRC open-weight AI by sending a new inquiry letter to DoorDash.
Congressional investigation focuses on risks of U.S. firms relying on competitively priced but potentially insecure Chinese AI models, including concerns over unauthorized model distillation and IP theft.
Lawmakers heightened scrutiny after the White House cited allegations that Moonshot, a Chinese AI firm, covertly distilled advanced U.S. models, intensifying worries about PRC-driven cybersecurity threats.

DoorDash request broadens congressional review

As reported by House Committee on Homeland Security, citing the House Committee on Homeland Security, Chairman Andrew R. Garbarino and House Select Committee on China Chairman John Moolenaar are continuing a joint investigation into how U.S. companies are evaluating and using open-weight AI models developed by the PRC.

The committees began the inquiry in April with letters to Anysphere and Airbnb, and say they have since held discussions with both companies on the technical, commercial, and security issues linked to those models. This week, the chairmen sent a similar letter to DoorDash seeking information as the probe moves into a new phase.

In the letter, the chairmen say U.S. businesses may be drawn to PRC-developed models because they offer competitive capabilities, lower costs, greater customization, and alternatives to a small group of proprietary model providers. They add that those advantages do not remove the need for risk-based safeguards or reduce concerns about growing dependence on models created by entities subject to PRC jurisdiction.

The investigation is examining where the models come from, how they were trained, whether their capabilities were improved through unauthorized distillation from leading U.S. systems, and how they are being repackaged or deployed in the United States and other markets. The committees are also assessing whether the United States has a sufficient open-weight AI strategy so domestic companies and cyber defenders are not forced to choose between expensive or restricted U.S. models and cheaper PRC-developed alternatives.

Security concerns sharpen focus on domestic AI capacity

Lawmakers are linking the inquiry to broader concerns that the PRC is intensifying efforts to narrow the AI gap with leading U.S. frontier labs through model distillation, intellectual property theft, and other illicit means. The update cites a July 22 statement by the director of the White House Office of Science and Technology Policy that Chinese AI company Moonshot may have run a covert platform for large-scale distillation against American AI models and trained systems using advanced technology it was not authorized to obtain.

The House Committee on Homeland Security says it continues to engage private-sector AI developers through roundtables, briefings, and hearings on cybersecurity and secure AI development. In May, Anthropic gave committee members a live demonstration of Mythos, which lawmakers say showed how advanced AI can identify software vulnerabilities and reinforced the need for federal agencies to use advanced U.S. models responsibly to find and patch security weaknesses.

Last week, members of both committees took part in a bipartisan war game focused on AI-enabled cyber threats, critical infrastructure resilience, and the homeland security implications of the PRC obtaining Mythos-level capabilities. In June, the Subcommittee on Cybersecurity and Infrastructure Protection also held a hearing on AI's effect on cybersecurity, infrastructure resilience, network security, and the strength of the domestic U.S. technology base.

The expansion of the Uyghur Forced Labor Prevention Act (UFLPA) Entity List was the focus of our earlier article, which detailed Washington’s move to add 43 more China-linked companies to a U.S. import blacklist tied to alleged forced labor in Xinjiang. We noted the broadened reach across sectors such as electronics, pharmaceuticals, metals, and cotton, and explained that the additions increase supply-chain enforcement pressure by raising the total number of listed entities to 187.

This material may contain third-party opinions, none of the data and information on this webpage constitutes investment advice according to our Disclaimer. While we adhere to strict Editorial Integrity, this post may contain references to products from our partners.