Sponsors (4)
Introduced March 26, 2026 by Donald Sternoff Beyer · Last progress March 26, 2026
The bill increases transparency, consumer access, and targeted safety disclosures for foundation models—helping buyers, researchers, and regulators—while imposing compliance costs and disclosure risks that could threaten privacy, competitiveness, and create regulatory uncertainty.
Users, downstream developers, and purchasers of AI models will get clearer information because developers must disclose training data sources, model capabilities and limitations, and safety measures.
Consumers, researchers, and regulators will have easier access to model information via a consumer‑friendly website and an FTC-hosted, machine‑readable central repository, making comparison and oversight simpler.
Small and new businesses will face lower short‑term burden to comply because the law provides tailored compliance help (grace periods, technical assistance, and templates).
Covered entities (including startups and established firms) will face new compliance costs for data collection, documentation, and reporting that could raise product prices, slow innovation, or divert resources.
Requiring disclosures about inference‑time data collection and retention could expose end users to privacy risks or create additional privacy‑compliance complexity.
Public disclosure requirements risk revealing proprietary model details or sensitive information (even with redaction allowances), which could harm commercial competitiveness or national/security interests.
Based on analysis of 2 sections of legislative text.
Requires the FTC to set rules forcing developers/providers of foundation models to submit and publish detailed model disclosures and safety/governance information in a central, machine- and human-readable repository.
Requires the Federal Trade Commission to write rules that force developers and providers of foundation AI models to publish and submit detailed information about those models. Within one year of enactment the FTC must, after consulting with technical agencies and stakeholders, set disclosure requirements for training data, data handling, model capabilities and limits, safety/monitoring practices, performance benchmarks, versioning, and other governance details, while allowing protection for sensitive data and option to publish information inside a system/model card. The rules must create a process for submitting required information to the FTC, make specified information publicly available in both human- and machine-readable formats (including a Commission-hosted central location), and include guidance to help covered entities comply. No new funding or specific enforcement provisions are created in the text provided; the FTC will use its rulemaking and enforcement authorities to implement the requirements.