HR 7209
TRAIN Act
TL;DR
The TRAIN Act, introduced by Rep. Madeleine Dean (D-PA), would let copyright holders subpoena AI companies to find out if their copyrighted works were used to train generative AI models. If an AI developer refuses to comply, the law would presume they used the copyrighted material without permission, making infringement lawsuits much easier to win.
How This Might Impact Your Business
Generative AI developers (think foundation model builders like those making LLMs, image generators, and code assistants) face a new legal discovery tool that lets rights holders demand training data disclosure.
Copyright holders including publishers, music labels, film studios, stock photo agencies, and individual artists gain a streamlined path to sue AI companies for unlicensed training data use.
Companies that fail to disclose training data when subpoenaed face a legal presumption of infringement, effectively flipping the burden of proof onto the AI developer.
Enterprises using third-party AI models (in marketing, product development, or customer service) could face indirect exposure if their vendors get hit with infringement claims and pull features or hike prices.
No company size threshold is specified, meaning startups and Big Tech alike would be subject to the same disclosure obligations.
The bill sits in the House Judiciary Committee with no hearing scheduled yet, so passage this session is uncertain but the concept has bipartisan interest.
Licensing markets for training data (news archives, image libraries, book catalogs) would likely expand as AI companies seek to preempt lawsuits with paid deals.
What Should You Do
Ask your legal and engineering teams to document what training data your internal AI models used and whether you have licenses or fair use defenses ready.
If you license AI tools from vendors (OpenAI, Anthropic, Google, etc.), review your contracts for indemnification clauses covering copyright infringement claims tied to training data.
Content owners should inventory copyrighted assets and consider whether to pursue licensing deals with AI developers now versus litigating later.
Track the bill through House Judiciary; watch for companion Senate legislation, which would signal real momentum.
Brief your board on training data risk as part of broader AI governance, especially if AI is central to your product roadmap.
Who It Affects
Sponsors
Status Timeline
committee
Referred to the House Committee on the Judiciary.
January 22, 2026