HR9341-119

What This Bill Does

The AI-Ready Federal Data Guidelines Act would require the National Institute of Standards and Technology to develop voluntary guidelines that help Federal agencies prepare datasets, including open Government data, for training AI models. NIST must consult the Office of Science and Technology Policy, the Department of Energy, the Office of Management and Budget, and other relevant agencies. The guidelines would address machine-interpretable formatting, data labeling and annotation, quality evaluation, metadata and documentation, ongoing maintenance, automated access to public data, flexible sector implementation, consistency with OMB Circular A-119, and conformity assessment.

NIST could run optional pilots to develop conformity-assessment procedures for specific scientific or industrial domains. Each pilot could last no more than one year, no more than two could run concurrently, and priority would go to areas with national-security and industrial-competitiveness implications, such as biotechnology and biomanufacturing, where Federal agencies hold relevant datasets. Pilots could use federally funded research programs, National Laboratories, universities, or private-sector partnerships. They would test data usability and interoperability, identify implementation costs and obstacles, and prepare for possible transfer of supplemental guidelines to a suitable non-Federal organization.

NIST must brief the House Science and Senate Commerce committees within one year after publishing the guidelines and annually for five years. It may not transfer or reprogram funds from another NIST program, office, project, or activity to carry out the new section. The bill also defines the covered agencies, AI models and systems, biomanufacturing, conformity assessment, universities, National Laboratories, and open Government data, and repeals a separate subsection of the NIST Act.

Who Benefits and How

Federal agency data managers could use a common voluntary framework to decide whether datasets are structured, documented, maintained, and accessible enough for AI use. AI developers, researchers, and public users could receive more usable and interoperable Government training data. National Laboratories, universities, federally funded research programs, and private technology partners could participate in pilots. Biotechnology and biomanufacturing researchers could receive early sector-specific guidance. Non-Federal standards organizations may eventually maintain supplemental pilot guidance. Other NIST programs benefit from the express prohibition on transferring their funds.

Who Bears the Burden and How

NIST staff must coordinate across agencies, write and maintain the guidelines, design any pilots, evaluate usability and resource needs, prepare transition materials, and brief Congress every year for five years. OSTP, Energy Department, OMB, and participating agency staff must provide technical and policy input. Pilot participants may incur data-formatting, annotation, quality, metadata, maintenance, and conformity-assessment costs. Agencies remain free not to adopt the guidelines, and NIST cannot pay for the work by reprogramming funds from other Institute activities, which may constrain implementation unless separate resources are available.

Key Provisions

• Requires NIST to develop voluntary, cross-agency guidelines for AI-ready Federal datasets and open Government data assets.
• Covers formatting, labeling, quality, metadata, maintenance, access, flexible use, OMB standards policy, and conformity-assessment procedures.
• Authorizes up to two concurrent one-year pilots through research programs, National Laboratories, universities, or private-sector partnerships.
• Prioritizes pilot topics with national-security and industrial-competitiveness importance, including biotechnology and biomanufacturing.
• Requires annual congressional briefings for five years after guideline publication and bars transfers from other NIST programs to fund the work.