READ AI Models Act
- Sponsor
- Rep. McBride, Sarah [D-DE-At Large]
- Committees
- Science, Space, and Technology Committee (primary)
- Last action
- Jun 25, 2026
Bottom line
The bill establishes a framework for NIST to create standardized documentation resources for AI models, promoting transparency and responsible AI development.
What it actually does
This bill directs the National Institute of Standards and Technology (NIST) to establish a pilot program for developing a structured template and technical guidelines for documenting artificial intelligence models and associated data. These resources are intended for use by both public and private sector entities, aiming to standardize information such as model developer, release date, training data knowledge cutoff, and terms of service. The template is designed to be modular and developed in consultation with various stakeholders.
Proponents argue
Supporters argue that standardized documentation is crucial for fostering responsible AI development, increasing trust, and providing essential information for evaluating AI systems. They contend that clear guidelines will promote interoperability, enhance accountability, and enable better understanding of how AI models function and their potential impacts.
Opponents contend
Critics might argue that mandating or even encouraging specific documentation standards could impose an undue administrative burden on AI developers, potentially stifling innovation in a rapidly evolving field. They might also suggest that a 'one-size-fits-all' template may not adequately address the diverse needs and complexities of various AI applications, leading to either over-documentation or insufficient detail in specific contexts.
The bill is concise and clearly written, making it accessible for a quick and informed evaluation.
Section 2(a)(1)
Directing NIST to Establish AI Model Documentation Pilot Program
This provision mandates the National Institute of Standards and Technology (NIST) to launch a pilot program. This program will create a standardized template and accompanying technical guidelines for documenting artificial intelligence models and their related data. These resources are intended for use by both public and private sector entities to enhance transparency and understanding of AI systems.
Supporters argue
Supporters contend this program is crucial for establishing a baseline of transparency and accountability in AI, enabling better understanding, evaluation, and responsible deployment of AI systems across various sectors. It provides a common language for discussing AI capabilities and limitations.
Critics contend
Critics might argue that creating a pilot program for documentation could lead to future regulatory burdens or that the resources might not keep pace with rapid AI advancements, potentially stifling innovation by adding compliance overhead.
Tradeoffs
This provision balances the need for greater transparency and accountability in AI with concerns about potential administrative burden on developers and the agility required in a fast-evolving technological field.
Section 2(a)(2)-(3)
Content and Design of the Structured Template and Technical Guidelines
This section details the specific requirements for the NIST-developed template and guidelines. The template must include information such as the model's name, developer, release date, knowledge cutoff for training data, supported languages, and terms of service. It must also be modular, allowing users to flexibly adopt sections based on specific use cases. The technical guidelines are to provide relevant metrics and benchmarks, incorporating voluntary consensus-based standards and industry best practices. NIST is also mandated to collaborate with private sector entities, academia, non-profits, international standards organizations, and federal agencies, and to solicit public comments on drafts.
This subsection defines 'artificial intelligence model' as a software component of an information system that implements artificial intelligence technology and uses computational, statistical, or machine-learning techniques to produce outputs from a defined set of inputs.
Section 2(a)(5)(C)
Why it matters:It is standard legislative practice to define key terms to ensure clarity and consistent application of the law, rather than embedding the definition within the main text.
Case for: This definition provides a clear and specific scope for what types of systems NIST's documentation requirements will apply to, preventing ambiguity and ensuring that the resources developed are targeted appropriately. It helps delineate the boundaries of the bill's impact.
Case against: Critics might argue that the definition, while specific, could be either too broad, potentially encompassing systems not typically considered 'AI' by some, or too narrow, potentially excluding emerging AI paradigms or specific applications that could benefit from documentation standards.
Estimated impact: The definition directly determines the precise scope of NIST's work in developing documentation resources and identifies the types of AI models that would fall under the proposed guidelines, thereby shaping the entire program's reach.