June 30, 2026 | Endorsements

H.R. 8283, Deterring American AI Model Theft Act of 2026

June 30, 2026 Endorsements

H.R. 8283, Deterring American AI Model Theft Act of 2026

Bottom Line Up Front

H.R. 8283, the Deterring American AI Model Theft Act of 2026, addresses the growing threat of Chinese AI firms conducting “model extraction attacks” against leading American AI companies. The bill directs the Secretary of State to assess which foreign entities of concern are extracting the capabilities of American closed-source AI models. It establishes a public “AI Model Extraction Attackers List,” and it makes perpetrators eligible for placement on the Commerce Department’s Entity List and blocking sanctions. It also closes a loophole through which China circumvents AI chip export controls by distilling U.S. closed-source models instead of training its own.

Bill Number H.R. 8283
Bill Name Deterring American AI Model Theft Act of 2026
Summary Directs an assessment of foreign model extraction attacks against American closed-source AI models, creates a public AI Model Extraction Attackers List, and authorizes Entity List placement and sanctions against perpetrators.
Chamber House
Lead Sponsor Rep. Bill Huizenga (R-MI)
Original Cosponsors Rep. John Moolenaar (R-MI)
Date Introduced 04/15/2026

Click here to read the full legislative text and view the list of cosponsors.

Why It Matters

  • Closing the distillation loophole: China is working to catch up on AI through distillation attacks, using large-scale queries to extract and replicate the capabilities of U.S. models built at great cost. This gives Beijing a shortcut to advanced AI capabilities and erodes a U.S. compute advantage built over years.
  • Export controls with consequences: Semiconductor manufacturing equipment controls and chip controls have limited the competitiveness of Chinese AI labs. Model extraction attacks let China circumvent those chip export controls by distilling U.S. closed-source models instead of building its own, and this bill imposes sanctions and penalties that make such attacks cost more than they yield.
  • Names and shames perpetrators: The bill directs the Secretary of State to assess which entities of concern are conducting model extraction attacks and to maintain a public AI Model Extraction Attackers List, with perpetrators made eligible for placement on the Commerce Department’s Entity List and blocking sanctions.
  • Protects legitimate research: Authorized distillation conducted with the model owner’s approval is a legitimate research method. The bill disentangles that lawful activity from illicit model extraction attacks, so it targets theft without burdening innovation.

“The United States is in a global AI arms race with the Chinese Communist Party. That’s why America’s cutting-edge chips and advanced technology should not power the AI systems that support China’s military and security state, especially at the expense of American innovators. FDD Action is proud to endorse the Deterring American AI Model Theft Act to prevent China from stealing its way to AI dominance.”

Connor Pfeiffer

Senior Director of Government Relations, FDD Action

Congressional Press Releases

Read statements from members of Congress who are supporting this legislation:

About FDD Action

FDD Action is a 501(c)(4) advocacy organization that works directly with policymakers to advocate for a robust U.S. foreign policy — one that strengthens U.S. national security, does damage to America’s adversaries, and supports allies and partners. FDD Action serves as a trusted resource for congressional offices and executive branch policymakers navigating complex national security challenges, leveraging a team with decades of experience on Capitol Hill and in the policy arena.

Issues:

China Export Controls