A PYMNTS Company

In Wake of Trump’s AI Action Plan, Senators Introduce Bipartisan AI Transparency Bill

 |  August 4, 2025

President Trump’s AI Action Plan did not have much to say about the data used to train AI models, but it remains a live topic on Capitol Hill. Two days after the Action Plan was released, a bipartisan group of senators reintroduced the Transparency and Responsibility for Artificial Intelligence Networks (TRAIN) Act, to provide rights owners a legal path to discover whether their works were used to train a generative AI model.

    Get the Full Story

    Complete the form to unlock this article and enjoy unlimited free access to all PYMNTS content — no additional logins required.

    yesSubscribe to our daily newsletter, PYMNTS Today.

    By completing this form, you agree to receive marketing communications from PYMNTS and to the sharing of your information with our sponsor, if applicable, in accordance with our Privacy Policy and Terms and Conditions.

    The bill would allow rights owners to ask a federal district court to issue a subpoena to an AI developer for information “sufficient to identify with certainty” whether their works were used. The rights holder would be required to submit an affidavit declaring a subjective good-faith belief that their works were used in a model, and that the subpoena’s purpose is to protect their rights.

    Failure to comply with the subpoena would create a presumption that the model developer made copies of the protected works.

    “This is simple: if your work is used to train AI, there should be a way for you, the copyright holder, to determine that it’s been used by a training model, and you should get compensated if it was,” the bill’s main sponsor, Sen. Peter Welch (D-VT), said in a news release. “As AI evolves and gets more embedded into our daily lives, we need to set a higher standard for transparency.”

    Co-sponsors of the measure include Sens. Marsha Blackburn (R-TN), Josh Hawley (R-MO), and Adam Schiff (D-CA).

    “AI should be in service to the American people—not the other way around. But under current law, Big Tech’s AI companies are stealing the works of today’s creators as they box out the next generation of creators,” Sen. Hawley said in a statement. “Congress should ensure that copyright holders can assert their rights against AI companies that are pirating creative works, and this bill gives Americans the tools to do so.”

    Read more: Tech Companies On Board With White House Plan to Ease Healthcare Record Keeping

    The TRAIN Act was originally introduced in 2024, but never came to a vote.

    The re-introduction was welcomed by creative industries groups in a press release issued by Sen. Welch’s office, including the Recording Industry Association of America, the National Music Publishers Association, the Authors Guild, SAG-AFTRA, the three main U.S. music performing rights organizations, and the American Society of Collective Rights Management representing photographers and graphic artists.

    How far the bill could go toward passage this time, however, is unclear. The president’s AI Action Plan’s one reference to data use calls for the creation of “the world’s largest and highest quality AI-ready scientific datasets,” for use in developing AI systems, and for promulgating regulations under the Confidential Information Act “on presumption of accessibility and expanding secure access, which will lower barriers and break down silos to accessing Federal data.”

    Otherwise, the Plan tilts heavily toward deregulating the AI sector. That includes denying federal AI funds to states with “burdensome AI regulations,” creating carveouts from federal environmental laws for AI data centers, and working with directing federal agencies to “identify, revise, or repeal regulations, rules, memoranda, administrative orders, guidance documents, policy statements, and interagency agreements that unnecessarily hinder AI development or deployment.”

    Whether data transparency laws such as the TRAIN Act would be regarded by the administration as unnecessarily hindering AI development remains to be seen.