Two major bills now advancing through Congress could reshape how creators protect their work in the age of artificial intelligence. H.R. 7209, the TRAIN Act, and S. 3813, the Copyright Labeling and Ethical AI Reporting (CLEAR) Act, represent the first serious legislative push to require transparency about which copyrighted works are used to train generative AI models. For creators, publishers, songwriters, and catalog owners, understanding these AI disclosure bills 2026 is no longer optional—it’s essential. This guide breaks down what these bills do, why they matter, and what you need to know to protect your work.
The Core Problem: Why AI Disclosure Bills Matter Now
Generative AI systems have been trained on billions of copyrighted works—often without permission, notification, or compensation to the creators who made them. Artists, writers, musicians, and photographers have filed lawsuits arguing that using their work to train AI violates their exclusive right to reproduce and license their creations. Meanwhile, AI developers argue that such training falls under fair use.
But the real issue for most creators isn’t abstract legal theory—it’s visibility. Right now, there is no reliable way for a songwriter to know if their lyrics were used in training data. A photographer has no clear path to discover whether their images were scraped and fed into an image-generation model. Publishers cannot audit whether their catalog was included. This information gap has made it nearly impossible for rights holders to make informed decisions about enforcement or licensing.
That’s where these two bills come in. Rather than settling compensation or liability questions directly, both focus on disclosure and transparency. They would require AI companies to report what copyrighted works they used—and create mechanisms for creators to find out.
H.R. 7209: The TRAIN Act and Administrative Subpoenas
H.R. 7209, formally titled the Transparency and Responsibility for Artificial Intelligence Networks Act (TRAIN Act), was introduced on January 22, 2026. The bill’s core mechanism is an administrative subpoena process that would allow copyright owners and their representatives to compel AI developers to disclose whether specific copyrighted works were used in training.
In practical terms, this means a songwriter, publisher, or administrator could issue a formal request to an AI company asking: “Was my work in your training data?” The company would then be required to respond with records and documentation. If they refuse or fail to comply, copyright owners could pursue enforcement action.
Why Administrative Subpoenas Matter
An administrative subpoena is not the same as a full lawsuit. It’s a lower-burden way to demand information before litigation becomes necessary. This is significant because it shifts the cost and complexity away from creators, who often lack the resources to sue major AI companies. Instead, they can request disclosure directly and use that information to decide whether to license, negotiate, or pursue legal action.
Music industry groups have already endorsed this approach. The National Music Publishers’ Association (NMPA) and BMI have both framed the TRAIN Act as essential leverage for songwriters and composers trying to audit use and evaluate enforcement or licensing options.
S. 3813: The CLEAR Act and Mandatory Notice Requirements
Here’s what the bill would require:
- Detailed summaries: AI companies must submit a detailed summary of every copyrighted work in their training dataset to the Register of Copyrights.
- Dataset URLs: If the training dataset is publicly available online, the company must provide the URL where it can be found.
- 30-day advance notice: Companies must file this notice at least 30 days before the AI model is commercially used—including any internal use within an organization.
- Public database: The Copyright Office would maintain a publicly searchable database of all submitted notices, making it possible for creators to search for their work.
Penalties and Enforcement
The CLEAR Act includes teeth. Companies that fail to file the required notice could face a private cause of action brought by copyright owners in federal district court. Penalties include:
- A minimum of $5,000 per instance of non-compliance
- Injunctive relief preventing the company from using that training data until notice is properly filed
- Civil penalties up to $2.5 million (without limiting other copyright remedies)
- Attorney’s fees and costs for copyright owners who prevail
These penalties are designed to make compliance the cheaper option than evasion.
How These Bills Work Together: A Transparency Framework
The two bills are complementary. The CLEAR Act creates a proactive disclosure regime—AI companies must report what they used before release. The TRAIN Act creates a reactive discovery mechanism—creators can demand information if they suspect their work was used and it wasn’t properly disclosed.
Together, they form what policy experts call a “disclosure-and-transparency framework” rather than a compensation framework. In other words, the immediate question is not “how much should creators be paid?” but rather “what must AI developers tell us about what they used?”
This distinction matters. If the bills pass, the practical burden shifts to AI companies to maintain detailed records, document their datasets, and file proper notice. For creators, it improves the ability to identify possible uses of their work and strengthens their negotiating position before disputes arise.
What This Means for Your Work
If either or both bills become law, you would gain:
- Visibility: A searchable public database where you can look up whether your work appears in AI training datasets.
- Information for negotiation: Proof of use becomes leverage. You can approach AI companies with evidence and negotiate licensing terms instead of discovering misuse after the fact.
- Audit rights: You gain a legal mechanism to compel disclosure if you suspect your work was used.
- Stronger enforcement options: Better documentation and provenance records make it easier to prove infringement if litigation becomes necessary.
What Creators Should Do Now
These bills are still in Congress, and their fate is uncertain. But the U.S. Copyright Office’s legislative tracker has now listed both bills as active measures, placing them alongside earlier AI-related bills from the 118th Congress—a signal that federal copyright policy is moving toward transparency requirements.
While you wait to see whether these bills pass, consider these practical steps:
- Document your work’s creation: Create timestamped evidence of when you made your work. This becomes crucial if you need to prove authorship or challenge unauthorized use in AI training. A notarized certificate of creation creates dated, witnessed proof that stands up as evidence.
- Register with the U.S. Copyright Office: If these bills pass and you need to enforce your rights, federal registration will be required before you can sue. Don’t wait until a problem arises.
- Monitor your work: Use available tools to search the web for unauthorized copies of your work. Some AI companies have already published their training datasets or provided tools to check whether specific works were included.
- Stay informed: Follow developments in these bills and similar legislation. The copyright landscape is shifting rapidly, and early awareness helps you stay ahead of new requirements or opportunities.
The Bigger Picture: What Happens Next
These bills represent a turning point. For years, the relationship between AI developers and creators has been opaque and adversarial. AI companies have trained on copyrighted works at massive scale, often without disclosure. Creators have had no reliable way to know or object.
Mandatory disclosure won’t solve all these problems, but it will change the playing field. Once creators can see what’s been used, they can make informed decisions. They can negotiate. They can enforce. They can build a more formal market for AI training licenses, with better documentation becoming a competitive requirement rather than an afterthought.
For now, the key is understanding that these bills are not about compensation—they’re about transparency and leverage. They assume that once creators have information, they can act. That’s why they matter so much.
Protecting Your Work While the Law Catches Up
Whether or not these bills pass, your responsibility as a creator remains the same: build strong, dated evidence of your authorship. The clearer your claim of creation, the stronger your position if disputes arise—whether over AI training, traditional copyright infringement, or any other use.
A notarized certificate of creation creates exactly that kind of evidence. It timestamps your work, witnesses your claim of authorship, and creates a record that’s hard to challenge later. It’s not a substitute for U.S. Copyright Office registration (which you’ll still need if you ever file suit), but it’s a complementary layer of protection that strengthens your overall position.
As AI disclosure bills 2026 move through Congress, the message to creators is clear: documentation, transparency, and provenance are about to become much more important. Start building that foundation now.
