Proposed Zero Draft for a Standard on Documentation of AI Datasets and AI Models

The U.S. Chamber of Commerce (“the Chamber”) appreciates the opportunity to submit comments to the National Institute of Standards and Technology ("NIST") regarding its " Extended Outline: Proposed Zero Draft for a Standard on Documentation of AI Datasets and AI Model.

The Chamber commends NIST’s commitment to developing science-based, stakeholder-informed standards through its “zero draft” process. This approach reflects a thoughtful and inclusive methodology that can lead to the creation of robust, voluntary standards grounded in industry expertise and practical implementation.

As outlined in the Chamber’s AI Principles[2], we strongly support the development of industry-led, consensus-based standards as a cornerstone of responsible digital innovation. Voluntary standards provide the flexibility needed to accommodate rapid technological advancement while ensuring accountability and public trust.

Further, the Chamber’s AI Commission Report [3]emphasizes the critical role of "soft law" mechanisms—such as standards and best practices—in shaping ethical and effective AI governance. These tools enable sector-specific guidance, foster innovation, and support global competitiveness. NIST’s leadership in convening government and industry to co-develop these pre-standards is essential to maintaining U.S. leadership in AI.

In this spirit, the Chamber offers the following comments and recommendations on the proposed zero draft to help ensure the framework is practical, scalable, and aligned with real-world needs.

      I.         General Feedback

We appreciate the strengths demonstrated in the draft and respectfully offer the following areas for further refinement and consideration:

a.    Audience Differentiation: The framework effectively identifies distinct documentation audiences and explicitly addresses the need for public-facing documentation, which could benefit from greater standardization.

b.    Balanced Trade-offs: The document thoughtfully considers documentation trade-offs and appropriately includes carve-outs for security, privacy, and proprietary information.

c.     Flexible Template Design: The model documentation template in Section 6.3 strikes the right balance—providing relevant information without over-specifying the format, allowing for broad applicability across sectors and use cases.

d.    Leverage Existing Frameworks and Standards: We encourage NIST to consider existing model card documentation practices as a foundational reference for its zero draft. In addition, existing favorable international work in standards, such as ISO 42001 can serve as a foundation for NIST’s efforts. Building on established approaches can enhance consistency and reduce implementation burdens.

e.    Promote Flexibility Through Less Prescriptive Language: To support innovation and accommodate evolving technologies, we recommend that NIST adopt less prescriptive language in its draft. This will allow organizations the flexibility to tailor documentation to their specific use cases and maturity levels.

II. Areas Requiring Clarification

Despite the Proposed Zero Draft's strengths, we believe the following areas would benefit from further clarification:

a.        Public Documentation vs. Artifacts Disconnect

While the draft states its focus is on public documentation, it places significant emphasis on documentation artifacts, which are typically more relevant for audit purposes than public reporting. The relationship between the artifacts discussed in earlier sections and the templates in Clause 6 should be clarified to avoid confusion.

b.       Dataset Documentation Ambiguity (Section 6.2)

Section 6.2 would benefit from greater clarity regarding its intended scope and whether the section is meant to apply to all training datasets associated with a model-as referenced in Section 6.3. This clarification would help prevent confusion and ensure consistent interpretation across stakeholders.

While we appreciate the draft’s recognition that proprietary information should be excluded, we suggest that NIST reconsider the inclusion of certain elements in Sections 6.2 and 6.3 due to their potential sensitivity and ambiguity. Specifically, we recommend omitting:

• Types of algorithms used
• Model size in parameters
• Hyperparameter configurations

Additionally, we urge NIST to consider the risk of adversarial attacks and compromising system integrity when considering transparency. A number of datasets contain sensitive user information, requiring careful handling to uphold privacy and data security commitments. Moreover, while NIST should seek to achieve its transparency goals, it should also balance pitfalls due to the volume and complexity regarding data used to train personalized data with disproportionate disclosure.

Finally, requiring granular documentation of training datasets could violate licensing and contractual terms and expose organizations to increased litigation risk, potentially hindering innovation and weakening U.S. competitiveness in AI. A balanced approach that prioritizes model-level transparency while respecting legal and operational constraints around datasets is essential.

c. Appendices

We are unable to provide substantive comments on the Appendices until the issues outlined above are addressed and clarified.

III.         Conclusion

The Chamber appreciates NIST’s leadership in advancing trustworthy AI through collaborative, science-based work to drive further standards development. As AI technologies continue to evolve and permeate every sector of the economy, it is imperative the United States develop strong standards which are grounded in operational realities. By incorporating the recommendations outlined above, NIST can help outline a standard that not only reflects technical rigor but also supports innovation, competitiveness, and public confidence in AI systems. We look forward to continued engagement and stand ready to support NIST in this important endeavor.

Sincerely,

Michael Richards
Executive Director
Chamber Technology Engagement Center
U.S. Chamber of Commerce

[1] Extended Outline: Proposed Zero Draft for a Standard on Documentation of AI Datasets and AI Models available a  https://www.nist.gov/artificial-intelligence/ai-research/nists-ai-standards-zero-drafts-pilot-project-accelerate

[2]U.S. Chamber of Commerce. “U.S. Chamber Releases Artificial Intelligence Principles.” U.S. Chamber of Commerce, 23 Sept. 2019, https://www.uschamber.com/regulations/us-chamber-releases-artificial-intelligence-principles.

[3]U.S. Chamber of Commerce.Artificial Intelligence Commission Report. 9 Mar. 2023, https://www.uschamber.com/technology/artificial-intelligence/artificial-intelligence-commission-report.

About the author

Michael Richards

Michael Richards is the executive director of policy at the Chamber's Center for Technology Engagement.

Read more