WeVote
Log In

Bill

GE

Bill

HR 10152

Open-Source AI Leadership Act

119th Congress Introduced by Gabe Evans

The bill directs the Secretary of Commerce to promote adoption of qualified U.S. open AI models in commerce while assessing and publicly reporting on risks posed by foreign adversa

Subcommittee Consideration and Mark-up Session Held
1
WeVote Research Nonpartisan
Bill Summary · HR 10152

Overview

  • Bill: HR 10152, Open-Source AI Leadership Act
  • Session: 119th Congress (introduced Aug 27, 2026)
  • Sponsor: Rep. Gabe Evans (Co-sponsor)
  • Main aim: Direct the Secretary of Commerce to support the adoption and use of qualified open artificial intelligence models in interstate and foreign commerce, with a focus on promoting American open-source AI models and assessing foreign adversary models.

Purpose and Intent

  • Promote the adoption and use of qualified open AI models developed in the United States.
  • Create a framework for coordinating government, private sector, and international partners to advance open AI models.
  • Identify and address barriers to adoption, establish policy guidance, and monitor adoption with a view toward competitiveness and national security.

Key Provisions and Changes

  1. Adoption and Use of Qualified Open Models (Section 2)

    • The Secretary of Commerce must take actions to support adoption and use of qualified open models in commerce.
    • Actions include:
      • Designating a single point of contact to coordinate efforts among the Department of Commerce, private sector entities (including startups), agencies, states, and qualified foreign partners.
      • Reviewing relevant programs and directing program heads to support adoption and use of qualified open models.
      • Identifying barriers to adoption and taking appropriate actions to address them.
      • Entering into agreements to support adoption with private sector entities, agencies, states, and qualified foreign partners.
      • Developing policy recommendations to facilitate adoption and use of qualified open models.
      • Identifying evidence-based approaches, best practices, frameworks, benchmarks, or other methods to monitor adoption and to assess foreign adversary models.

  2. Assessment of Foreign Adversary Models (Section 3)

    • The Secretary must identify, assess, and publicly disclose risks related to foreign adversary models, including:
      • Risks from training foreign adversary models (including via unauthorized use of AI models) by covered nations or entities under their control.
      • Risks to confidentiality, integrity, and accessibility of personal and proprietary information.
      • Security and resiliency risks to organizations and their supply chains from using foreign adversary models.
      • Quality and implications of outputs from foreign adversary models (e.g., historical accuracy, scientific integrity, freedom of expression).
      • Effectiveness of safety or misuse-prevention features in foreign adversary models.
      • Risks related to development and use of chemical, biological, radiological, nuclear, or other security-sensitive capabilities.
    • Public reporting: The Secretary must publish a risk assessment report on a publicly accessible website within 18 months after enactment and annually thereafter, comparing foreign adversary models with qualified open models in terms of cost, capability, and performance.
    • Legislative reporting: Not later than 30 days after each public report, the Secretary must submit a report to the specified congressional committees.

  3. Sunset Provision (Section 3)

    • The annual public reporting requirement on foreign adversary models terminates 10 years after enactment.

  4. Rule of Construction (Section 4)

    • The Act does not authorize the Secretary to ban, restrict, or prohibit open AI models in interstate or foreign commerce.

  5. Definitions (Section 5)

    • Key terms defined, including:
      • Agency, Artificial Intelligence, Artificial Intelligence Model.
      • Covered Nation: defined per 10 U.S.C. 4872(f).
      • Foreign Adversary Model: an open AI model developed or distributed by a covered nation or entities under the control of a covered nation.
      • Open Artificial Intelligence Model: weights publicly released or distributed under an open license permitting use, modification, and redistribution.
      • Qualified Open Model: an open model developed and made available by a U.S. person in commerce and not developed or made available by a covered nation or entities under its control.
      • Qualified Foreign Partner: a foreign country or union not a covered nation.
      • U.S. Person: individuals/entities meeting standard residency, headquarters, and incorporation criteria.
      • Secretary, State, and related terms as used in the bill.

Who Would Be Affected

  • U.S. Department of Commerce and its agencies
  • U.S. and foreign private sector entities, including startups, engaged in open AI model development or deployment
  • States and potential foreign partners aligned with the act’s framework
  • Entities subject to foreign adversary model risk assessments (including those relying on or evaluating foreign open AI models)

Procedural and Timeline Aspects

  • Immediate effect: Establishes a framework for coordination, policy development, and barrier assessment.
  • Reporting timeline:
    • Public risk assessment report due within 18 months after enactment, and annually thereafter.
    • Follow-up congressional report within 30 days after each public report.
  • Sunset: Public reporting requirement ends 10 years after enactment.
  • Relationship to existing law: Does not authorize banning open models; focuses on adoption support and risk assessment.

Potential Impacts and Considerations

  • Encourages domestic development and proliferation of qualified open AI models.
  • Aims to increase transparency about risks associated with foreign adversary models and provide cost/capability comparisons.
  • Could influence procurement, partnerships, and standards for open AI in the U.S. by promoting open-source licensing and domestic leadership.
  • Balances openness with national security by requiring ongoing risk assessments and public disclosure, without imposing a ban.

Compiled from official sources — confirm details with the bill’s official record.

Sign in to ask a question.