- DevelopersMay lower litigation risk for AI developers when their products are used by licensed professionals, potentially reducin…
- Federal agenciesCreates explicit federal standards for transparency (model cards and model specifications), which could improve informa…
- DevelopersBy limiting developer liability (except for recklessness/willful misconduct) and allowing defined redactions for trade…
RISE Act of 2025
Read twice and referred to the Committee on Commerce, Science, and Transportation.
The Responsible Innovation and Safe Expertise (RISE) Act of 2025 creates a conditional federal safe harbor that shields certain artificial intelligence (AI) developers from civil liability for errors produced by their AI when those systems are used by licensed, "learned professionals" in providing professional services. To qualify for immunity a developer must publish and maintain a model card and a model specification (with limited, justified trade-secret redactions allowed), provide clear documentation to professionals about limitations and failure modes, and update those materials within 30 days after deploying a new version or discovering a new material failure mode.
Scope of liability: liberals see broad immunity as undermining accountability; conservatives see immunity as necessary for innovation.
Relative to its intended legislative type, this bill establishes a clear substantive legal change (conditional immunity tied to transparency obligations) with reasonably well‑specified core elements and definitions, but it provides limited implementation, enforcement, and fiscal detail.
The Responsible Innovation and Safe Expertise (RISE) Act of 2025 creates a conditional federal safe harbor that shields certain artificial intelligence (AI) developers from civil liability for errors produced by their AI when those systems are used by licensed, "learned professionals" in providing professional services.
To qualify for immunity a developer must publish and maintain a model card and a model specification (with limited, justified trade-secret redactions allowed), provide clear documentation to professionals about limitations and failure modes, and update those materials within 30 days after deploying a new version or discovering a new material failure mode.
The immunity excludes conduct that is reckless or willful, does not apply to fraud or knowing misrepresentation or non‑professional uses, and expressly preempts state-law claims against developers that fall within the federal safe harbor.
On content alone, the bill is a focused, administrable proposal that includes compromise features (transparency requirements, update duties, and exceptions). Nevertheless, it alters liability regimes and preempts state law — areas that historically trigger strong stakeholder opposition and legal scrutiny. Without clear, broad consensus among diverse affected constituencies (healthcare, legal, finance, consumer groups, state regulators) or bundling into a larger negotiated legislative vehicle, its standalone path to becoming law is uncertain.
Relative to its intended legislative type, this bill establishes a clear substantive legal change (conditional immunity tied to transparency obligations) with reasonably well‑specified core elements and definitions, but it provides limited implementation, enforcement, and fiscal detail.
Scope of liability: liberals see broad immunity as undermining accountability; conservatives see immunity as necessary for innovation.
Who stands to gain, and who may push back.
These are examples from the analysis, not a ranked list of the most-affected groups.
- DevelopersCould reduce legal recourse for clients injured by AI errors because developers who meet the disclosure conditions are…
- DevelopersMay increase compliance costs for developers (preparing, publishing, updating model cards/specifications, and documenti…
- DevelopersAllowing redactions for trade secrets with only a developer’s contemporaneous justification could limit meaningful tran…
Why the argument around this bill splits.
Scope of liability: liberals see broad immunity as undermining accountability; conservatives see immunity as necessary for innovation.
A mainstream progressive would view this bill skeptically.
They would acknowledge the value of requiring public model cards and specifications, but be concerned that the immunity and federal preemption could unduly limit harmed clients’ ability to obtain redress and weaken incentives for developers to make their systems safer.
Trade-secret redaction allowances, the 30-day update window, and the preemption of state claims would be seen as gaps that could protect developers at the expense of client protections and accountability.
A pragmatic moderate would see the bill as a reasonable attempt to balance legal certainty for AI developers with transparency obligations to professional users, but would note several ambiguities that merit fixes.
They would appreciate the model card/specification and update requirements as mechanisms that encourage responsible deployment, while worrying that federal preemption and the immunity standard (excluding only reckless/willful conduct) may be too protective of developers.
A centrist would likely support the bill if amended to tighten definitions, strengthen update enforcement, and preserve certain state remedies or provide federal procedural safeguards.
A mainstream conservative would generally view the bill favorably because it reduces developer liability risk and creates a uniform federal standard, which promotes innovation and reduces exposure to disruptive state-by-state litigation.
The conditional nature of the safe harbor (tying immunity to transparency and updates) is likely seen as a reasonable market‑oriented compromise that preserves incentives for developers to provide useful information to professional users.
Some conservatives might still want stronger protection for proprietary information and less regulatory burden in the disclosure and update obligations.
The path through Congress.
Reached or meaningfully advanced
Reached or meaningfully advanced
Still ahead
Still ahead
Still ahead
On content alone, the bill is a focused, administrable proposal that includes compromise features (transparency requirements, update duties, and exceptions). Nevertheless, it alters liability regimes and preempts state law — areas that historically trigger strong stakeholder opposition and legal scrutiny. Without clear, broad consensus among diverse affected constituencies (healthcare, legal, finance, consumer groups, state regulators) or bundling into a larger negotiated legislative vehicle, its standalone path to becoming law is uncertain.
- What specific standards or enforcement mechanisms will determine whether a model card or model specification meets the Act's requirements — the bill references "industry standards" but does not define who adjudicates sufficiency.
- How courts will interpret the scope of preemption and exceptions (e.g., what constitutes "conduct outside the scope of professional use" or when a redaction is inappropriate), which could lead to litigation challenging the statute or narrowing its effect.
Recent votes on the bill.
No vote history yet
The bill has not accumulated any surfaced votes yet.
Go deeper than the headline read.
Scope of liability: liberals see broad immunity as undermining accountability; conservatives see immunity as necessary for innovation.
On content alone, the bill is a focused, administrable proposal that includes compromise features (transparency requirements, update duties…
Relative to its intended legislative type, this bill establishes a clear substantive legal change (conditional immunity tied to transparency obligations) with reasonably well‑specified core elements and definitions, but…
Go beyond the headline summary with full stakeholder mapping, legislative design analysis, passage barriers, and lens-by-lens tradeoff breakdowns.