Copyright owners get new power to demand AI training data—with a low bar.
H.R. 7209 — TRAIN Act · Filed by Madeleine Dean (D-PA) · 3 cosponsors · Introduced Jan 22, 2026 · Referred to committee
Your members of Congress
Enter a ZIP to see where your representative and both senators stood on this bill.
Looked up on this device — your ZIP is never stored on our servers.
What it does
This bill creates a new legal tool allowing copyright owners to subpoena AI developers and demand disclosure of which copyrighted works were used to train their AI models. A copyright owner files a sworn declaration with a court clerk stating they believe their work was used in training, and the clerk must issue the subpoena if the paperwork is in order. The developer must then disclose the training materials or records identifying them. Failure to comply creates a legal presumption that the developer did use the copyrighted work.
Why we flagged it
The bill's core function is to create a new subpoena mechanism for copyright owners to compel disclosure of AI training data. It is not a ban or prohibition on AI training, but rather a procedural right to discover whether infringement occurred. The mechanism is narrow and targeted at the copyright dispute context.
What the text implies
- The 'subjective good faith belief' standard is extremely low and does not require evidence before subpoena issuance—a copyright owner need only declare they believe their work was used, and the clerk must issue the subpoena. This may enable fishing expeditions and harassment of AI developers.
- The rebuttable presumption of infringement upon non-compliance shifts the burden to the developer to prove they did NOT use the work, reversing the normal copyright burden of proof and creating strong incentive to comply even with questionable subpoenas.
The full analysis lists 5 implications of this text.
Who stands to gain
copyright owners and publishers; entertainment and media companies; literary and visual artists