Senator Josh Hawley challenges YouTube executive Suzana Carlos

Media and entertainment personalities worried that their work is being unlawfully fed into AI copyright models gained a prominent ally Wednesday when Missouri Senator Josh Hawley went after a YouTube executive for how Google trains its models in a tense Capitol Hill session.

“That seems like a big problem to me,” Hawley said in a Judiciary subcommittee hearing after the executive, Music Policy chief Suzana Carlos, had explained that videos uploaded to the site are used to train Gemini and other Google AI products for any user agreeing to the terms of service, a requirement for uploading. “That seems like a huge, huge problem to me. And the fact that YouTube is monetizing these kinds of videos seems like a huge problem to me.”

Hawley said that Congress and tech companies needed to find ways to “give individuals powerful enforceable rights and their image and their property and their lives back again.”

The Republican lawmaker was speaking at a hearing for the No Fakes Act, where earlier in the session Carlos, country star Martina McBride and RIAA chief executive Mitch Glazier had been among those testifying on behalf of the bill, which aims to impose limits on how users might create AI versions of an artist or other person’s face or voice. The previous two iterations of the bill had not gained enough support to pass, but backers hope YouTube coming aboard, as it recently did, could spell the difference.

But it was Hawley’s remarks on an only partly-related issue that were among the most notable at the hearing. AI copyright has been a key battle area between media and tech companies, with The New York Times currently suing OpenAI over how models are trained on its articles in a closely watched case for anyone who creates or holds intellectual property. Tech companies need the content company’s material to build their models, which rely on tens of thousands of news stories, images, videos, songs and other content to generate their output.

Trump’s removal of the head of the U.S. Copyright Office earlier this month after she expressed hesitation that what OpenAI and others were doing is legal has further stirred concern that copyrighted material is being used unlawfully without artists’ and companies’ consent.

Hawley has previously positioned himself as an anti Big Tech-populist, going after Meta for a host of alleged missteps and, more specifically on AI, urging the Labor Department to protect workers in the wake of AI shifts and criticizing OpenAI for its accelerationism.

But Wednesday marks some of his strongest public words yet on the copyright issue, and makes the Republican lawmaker something of a strange bedfellow to SAG-AFTRA, Justine Bateman, Hollywood artist Reid Southen, The New York Times and others in creative fields who have raised alarm bells that a business is being built on the backs of others’ work without compensation.

“We’ve got to do more. YouTube is I’m sure making billions of dollars off this,” Hawley said Wednesday. “The people losing are the artists and creators and the teenagers whose lives are upended.” (The last point was a reference to deepfakes involving bullying and non-consensual images, which Hawley also said YouTube needs to police harder.)

For the most part tech companies have said the material they train their models on is allowed under laws of fair use. Carlos repeated several times during Hawley’s questioning that Google trains its models on videos uploaded to YouTube. “We do share data in accordance with our agreements,” she said.

Hawley said he believed that none of the major tech companies have sufficiently explained how they’re protecting a creator’s copyright and that more of them need to answer for their training actions.

Hawley criticized those firms too. “I’m glad you’re here today,” he told Carlos. “I wish there were more tech companies here today.”