RISE Act of 2025
Sponsored By: Senator Lummis, Cynthia M. [R-WY]
Introduced
Summary
Conditional civil immunity for AI developers is the core change in this bill. It would create a federal framework that allows developers to avoid civil liability when their AI products are used by licensed professionals, while defining key terms and documentation requirements.
Show full summary
- Developers: The bill would let a developer be immune from civil suits for certain errors produced by an AI product when the product is used by a learned professional in providing services. It also defines a developer to include creators, designers, and entities that market or control an AI product.
- Learned professionals: A learned professional is defined as a licensed or authorized practitioner who is bound by professional standards and must exercise independent professional judgment when using AI. That definition is a condition tied to developer immunity.
- Clients: The bill clarifies who counts as a client by tying the term to people who engage and rely on a professional under standards or regulations.
- Transparency and technical disclosure: The bill requires publicly available model cards that describe training data, evaluation methods, performance, intended uses, limitations, and risk mitigations. It also defines model specifications to include system prompts, hidden configuration, training guidance, and instructions given to human raters.
Personalized for You
How does this bill affect your finances?
Personalize government policy and PRIA will tell you what this bill means for your household, plus every other piece of legislation we track. PRIA reads each provision against your financial profile to show you exactly what matters to your wallet.
Bill Overview
Analyzed Economic Effects
1 provisions identified: 1 benefits, 0 costs, 0 mixed.
Safe harbor for AI developers
If enacted, this bill would give limited legal protection to AI developers when licensed professionals use their AI. To qualify, a developer would need to publish and keep a public model card and a model specification before deployment. Redactions would be allowed only for trade secrets unrelated to safety and must include a written justification. Developers must give clear documentation to professionals about limitations, failure modes, and safe uses. Immunity would not cover recklessness, willful misconduct, fraud, knowing misrepresentation, or uses outside professional services. State-law claims would be preempted only when a developer meets the safe-harbor conditions. Developers must update the model card, model specification, and documentation within 30 days after a new version or a newly found harmful failure or they would lose immunity for harms after that date. The bill would also define key terms like "artificial intelligence," "developer," "client," "learned professional," "error," "model card," and "model specification." If enacted, these rules would take effect on December 1, 2025.
Sponsors & CoSponsors
Sponsor
Lummis, Cynthia M. [R-WY]
WY • R
Cosponsors
There are no cosponsors for this bill.
Roll Call Votes
No roll call votes available for this bill.
View on Congress.gov