The AI Foundation Model Transparency Act of 2026 requires the Federal Trade Commission to create rules forcing companies that develop large artificial intelligence foundation models to publicly disclose detailed information about their training data, algorithms, and performance capabilities within one year of the bill's enactment. Companies subject to the law must provide information such as where training data comes from, what languages the model supports, known risks and limitations, performance on safety benchmarks (including handling of medical, national security, election, and hiring-related questions), and computational power used for training—with some sensitive information kept confidential to protect security and privacy. The rules apply to major AI developers meeting certain thresholds, such as those with over 10 million monthly users or models trained on extremely high computing power (10^26 operations), though fully open-source models are exempt and small businesses receive a three-month grace period with compliance assistance. The FTC will enforce these requirements as violations of consumer protection law, and it must update the regulations annually as the technology evolves.
Take Action
Your position
Add a comment
to comment on this bill.
Annotate the text
Highlight any passage on the Summary or Full Text tab to attach a note. Annotations appear on the Annotations tab.