U.S. Reps. Andrew Garbarino (R-NY) and John Moolenaar (R-MI) announced the launch of a joint investigation into the national security risks posed by artificial intelligence models developed by Chinese companies.
Garbarino, chair of the House Committee on Homeland Security, said his committee and the House Select Committee on China, of which Moolenaar is chair, will look into national security and cybersecurity risks adoption of AI models like DeepSeek, Alibaba, Moonshot AI, and MiniMax, pose to the United States. The investigation comes amid growing concern that Chinese companies may be using the AI models to extract capabilities from leading American frontier models, and then repackaging those capabilities into lower-cost models without the same safeguards as American ones.
As part of their investigation, the law makers sent letters to Anysphere and Airbnb asking for their concerns about the companies’ use of or exposure to the risks through PRC-developed AI. The letters followed a memo from the White House Office of Science and Technology Policy warning that foreign entities, primarily in China, are conducting deliberate, industrial-scale campaigns to distill U.S. AI systems through proxy accounts and other coordinated methods.
“The billions of dollars American companies invest in foundational research, compute infrastructure, and security engineering is being undercut by a sustained extraction campaign conducted at a fraction of the cost of independent development,” the two Congressmen wrote in a letter to Anysphere. “This threat is not limited to commercial harm. American frontier AI laboratories invest heavily in security testing and in building guardrails designed to prevent their models from being used to develop weapons, automate software vulnerability discovery and exploitation, generate tailored disinformation, or assist in the synthesis of dangerous chemical or biological agents. When capabilities are stripped out through distillation and repackaged without equivalent safeguards, the resulting models may become available to hostile state actors, terrorist organizations, and criminal enterprises.”
The chairmen said the PRC-developed open-weight AI have experienced rapid global adoption, and have grown to 30 percent of the global AI workloads, and that those models have been reported to exhibit censorship aligned with the Chinese Communist Party’s position on politically sensitive topics.
“Beijing has been explicit in its view that the global distribution of open-weight AI serves PRC strategic interests,” the leaders said. “What is at issue, therefore, is not simply market competition, but the growing risk that software systems used across the American economy, government, and defense industrial base will come to depend on models developed by PRC-linked laboratories and shaped by PRC strategic objectives.”