Directs the Federal Trade Commission to issue transparency regulations for AI model training data and algorithms in consultation with relevant stakeholders within 9 months. Specifies the information categories to be publicly disclosed and updated annually.
Analysis summaries, actor details, and coverage mappings were LLM-classified and may contain errors.
This is a binding legislative act from the United States Congress that directs the Federal Trade Commission to promulgate regulations with mandatory compliance requirements and enforcement mechanisms including penalties for violations.
The document has minimal to good coverage of approximately 6-8 subdomains, with primary focus on transparency and governance (7.4), privacy concerns (2.1), discrimination risks (1.1, 1.3), misinformation (3.1), and governance structures (6.5). Coverage is concentrated in system safety, transparency, and fairness domains, with the document primarily addressing these risks through mandatory disclosure requirements.
This is an external regulation that applies broadly across all sectors where foundation models are deployed. The document explicitly references multiple high-risk application domains including healthcare, finance, education, employment, public services, and law enforcement, indicating cross-sectoral governance of AI foundation models.
The document primarily addresses the Build and Use Model, Verify and Validate, Deploy, and Operate and Monitor stages. It focuses extensively on transparency requirements for training data and model documentation (Build), performance evaluation (Verify), deployment disclosure (Deploy), and ongoing monitoring through annual updates (Operate and Monitor).
The document explicitly defines and focuses on 'foundation models' as AI models trained on broad data with at least 1 billion parameters applicable across wide contexts. It does not explicitly mention frontier AI, general purpose AI, task-specific AI, generative AI, predictive AI, or open-weight models, though it does consider alternative provisions for open-source foundation models.
United States Congress; Mr. Beyer; Ms. Eshoo
The document is a bill introduced in the House of Representatives by Mr. Beyer and Ms. Eshoo, making the United States Congress the proposing entity.
Federal Trade Commission (Commission)
The Federal Trade Commission is explicitly designated as the enforcement authority with powers to promulgate regulations, enforce compliance, and impose penalties for violations.
Federal Trade Commission (Commission); National Institute of Standards and Technology; Office of Science and Technology Policy; Register of Copyrights
The FTC is designated to monitor compliance through submission requirements and annual assessments. NIST, OSTP, and the Register of Copyrights serve consultative monitoring roles in assessing and updating standards.
Covered entities (persons, partnerships, or corporations providing foundation models with over 100,000 monthly output instances or over 30,000 monthly users); Common carriers subject to Communications Act; Non-profit organizations
The Act targets 'covered entities' which are defined as persons, partnerships, or corporations that provide foundation models meeting specified usage thresholds, including those over which the FTC has jurisdiction, common carriers, and non-profit organizations.
15 subdomains (2 Good, 13 Minimal)