HR-9341-119
Ordered to be Reported (Amended) by the Yeas and Nays: 29 - 0.
Sponsored by Brian Babin (R-TX)
What it does
This bill would require the Director of the National Institute of Standards and Technology (NIST) to develop voluntary guidelines helping federal agencies prepare their open government datasets for use in training artificial intelligence models. The guidelines would cover data formatting, labeling, quality evaluation, metadata, maintenance, and availability. The bill also authorizes up to two concurrent, time-limited pilot programs (no longer than one year each) to test sector-specific guidelines, and requires NIST to brief Congress annually for five years on implementation progress.
Who benefits
AI developers and technology companies that would gain better-structured, more accessible federal datasets for training AI models. Federal agencies that would receive standardized guidance for managing and publishing data. Researchers and academics at universities and national laboratories who could participate in pilot programs and access higher-quality public data. Industries with national security and competitiveness implications — such as biotechnology and biomanufacturing — that are prioritized for pilot programs. Taxpayers broadly, if improved AI tools built on federal data lead to more efficient government services.
Who is hurt
Private data vendors and companies that currently sell curated or structured datasets to AI developers, who may face increased competition from improved free federal data. Federal agency staff who would bear implementation workload even though the guidelines are voluntary. NIST itself, which is explicitly prohibited from reprogramming funds from other programs to carry out this mandate, potentially straining existing resources. Smaller AI startups that lack the capacity to engage with conformity assessment procedures may be disadvantaged relative to larger firms better positioned to shape or comply with emerging standards.
Supporters argue
Supporters argue that the federal government holds vast troves of high-value public data — from health records to climate measurements — that are currently poorly structured for AI use, leaving American AI developers at a disadvantage relative to foreign competitors with better-organized national datasets. They contend that voluntary, NIST-led standards have a proven track record of driving industry adoption without regulatory burden, as demonstrated by the widely adopted NIST Cybersecurity Framework, and that this bill follows the same low-friction model. Prioritizing sectors like biotechnology and biomanufacturing directly addresses national competitiveness concerns identified in the CHIPS and Science Act.
Opponents argue
Opponents argue that the bill creates an unfunded mandate by prohibiting NIST from reprogramming existing funds while assigning a significant new workload, potentially degrading NIST's capacity in other critical areas without new appropriations. They contend that voluntary guidelines historically see uneven adoption across agencies, meaning the bill may produce standards that sit unused — a pattern seen with earlier open data mandates under the OPEN Government Data Act — without enforcement mechanisms to ensure agencies actually prepare AI-ready datasets. Critics may also raise concerns that improving automated access to federal datasets could inadvertently facilitate privacy risks if agencies lack the capacity to properly vet data before publication.