The Open-Source AI Leadership Act is a bill in the 119th Congress directing the Secretary of Commerce to support the adoption and use of American open-weight AI models, and to identify, assess, and publish the risks of open models developed by or under the control of covered nations. It was introduced in the House of Representatives on August 27, 2026 by Representative Gabe Evans (R-Colo.) and referred to the Committee on Energy and Commerce (H.R. 10152 — Open-Source AI Leadership Act (bill text)).
Status
| Date | Action |
|---|---|
| 2026-08-27 | Introduced in House by Rep. Evans of Colorado |
| 2026-08-27 | Referred to the House Committee on Energy and Commerce |
| 2026-09-01 | Approved at markup by the House Energy and Commerce Subcommittee on Commerce, Manufacturing, and Trade, chaired by Rep. Gus Bilirakis (R-Fla.), as one of twelve bills |
The subcommittee approved the bill on September 1, 2026 in a markup of twelve measures that also included H.R. 10170, the Memory Chip Competitiveness Assessment Act (Reps. Mariannette Miller-Meeks and Darren Soto); H.R. 6207, the Chip EQUIP Act (Reps. Zoe Lofgren and Jay Obernolte), with an amendment in the nature of a substitute; and H.R. 10163, the American Quantum Competitiveness Act (Rep. Nick Langworthy). Inside AI Policy characterized the package as addressing competitive and security challenges from China; only that account's lede was retrievable (Source: insideaipolicy.com; energycommerce.house.gov).
Scope and definitions
An open artificial intelligence model is defined as a model with weights publicly released for download or distribution, or one distributed under an open license permitting use, modification, and redistribution of source code and weights. A qualified open model is an open model developed and made available in interstate or foreign commerce by a U.S. person — a person domiciled in, headquartered in, and organized under the laws of the United States — and not developed or made available by a covered nation or an entity or individual under a covered nation's control.
A foreign adversary model is the mirror image: an open model developed or made available by a covered nation, or by an entity or individual subject to a covered nation's control. Covered nation takes the meaning in 10 U.S.C. § 4872(f), the statutory list used in defense acquisition. A qualified foreign partner is a foreign country or political and economic union that is not a covered nation. "Artificial intelligence" takes the definition in section 5002 of the National Artificial Intelligence Initiative Act of 2020 (15 U.S.C. 9401).
Key provisions
Support for qualified open models (Sec. 2)
The Secretary of Commerce is directed to take any of six actions as appropriate: designate a single point of contact to coordinate work between the Department, private-sector entities, and other stakeholders; review the Department's existing programs and powers and direct program heads to support adoption; identify and address barriers to adoption; enter into agreements with private-sector entities including startups, agencies, states, and qualified foreign partners; develop policy and recommendations to facilitate adoption; and identify evidence-based approaches, best practices, frameworks, benchmarks, or other methods for monitoring the adoption of both qualified open models and foreign adversary models.
Assessment of foreign adversary models (Sec. 3)
The Secretary must identify, assess, and make publicly known any risk related to adoption and use of foreign adversary models, including risks concerning: how such a model was trained, including training conducted through unauthorized use of another AI model; the confidentiality, integrity, and accessibility of personal and proprietary information, including access by a covered nation; the security and resiliency of organizations that adopt such a model and consequent supply-chain risk; model outputs, "including with respect to historical accuracy, scientific inquiry, and freedom of expression"; the efficacy of any safety or security feature intended to prevent misuse; and chemical, biological, radiological, or nuclear capabilities and other risks to national and economic security.
Not later than 18 months after enactment and annually thereafter, the Secretary must publish a report on a public website covering the identified risks, an overview of the adoption and use of leading foreign adversary models, their cost, capability, and performance, and a comparison of that information against the corresponding information for qualified open models. The report goes to the House Committee on Energy and Commerce and the Senate Committee on Commerce, Science, and Transportation within 30 days of publication. The reporting requirement sunsets 10 years after enactment.
Rule of construction (Sec. 4)
Nothing in the Act may be construed to authorize the Secretary "to ban, restrict, or otherwise make unavailable an open artificial intelligence model in interstate or foreign commerce." The provision bounds the bill to promotion, disclosure, and comparative assessment rather than availability controls, which distinguishes it from the licensing and distribution-control approaches discussed in Open-Source AI / Open-Weight Models.
Relationships
- depends-on: H.R. 10152 — Open-Source AI Leadership Act (bill text) — the Introduced-in-House bill text this page summarizes
- instance-of: Open-Source AI / Open-Weight Models — federal promotion of open-weight models as policy instrument
- related: Export Controls (AI) — addresses foreign adversary models through disclosure and assessment rather than trade controls
- related: US-China AI Competition: Different Races, Different Metrics
- related: AI Transparency
- related: AI Advertising Disclosure Act (H.R. 10146, 119th Congress) — introduced in the same week and referred to the same committee