As artificial intelligence draws increased attention, a recording industry leader provides perspective.

SoundExchange President and CEO, Michael Huppe.

© Elliott O'Donovan

AI’s presence in popular music is raising alarm with concerns about the cloning of copyrighted material and the technology’s potential replacement of artistic creativity at the press of a button.

In March 2026, IFPI (the International Federation of the Phonographic Industry) referenced the active engagement of record companies with AI, aiming to achieve a blend between the technology and human artistry. This includes the major label signing of licensing agreements with AI music generator Udio (except for Sony which is still litigating), music catalog companies, and other licensors. Unlike Udio which proposes a “walled garden” approach (a closed ecosystem that prevents users from uploading music from its platform), its rival, Suno, does not yet impose any such limitations.

When asked by Forbes about the AI challenges confronting the recording industry, Michael Huppe, the President and CEO of SoundExchange, considered the possibilities.

At this critical moment, how great is the AI threat to the recording industry’s intellectual property?

AI has the potential to be a threat and also a great opportunity. The threat of uncontrolled AI is very real. The advancements that are being made in AI technology every couple of weeks are remarkable and the output is becoming better and better. It is critical that our culture and our legal systems put in place the guardrails that are necessary to protect creators, especially human creators. I talk a lot about the three C's. It's okay to use artist content and to use creators content as long as you have consent, you credit them, and you compensate them.

Under that rubric, done properly with the proper safeguards by ethical companies who care about these things, AI has a lot to offer to the industry. Artists, producers, labels, and a lot of people in the music industry are making very positive use of AI. Rails need to be in place so that human creativity and human input are protected, recognizing the value that humans bring to the fore.

What needs to happen at a legislative level in America?

Well, let's talk legislative and let's also talk commercial. We would be remiss if we didn't say that licensing is happening. There are many AI companies out there, many AI deals being done with record labels and news organizations, et cetera, where you have the marketplace and content owners leaning in, trying to figure out how to work with these AI companies to create a path forward. But on the legislative front, we think copyright law in the U.S. gives us the protection we need now in terms of dealing with copyright infringement.

In the content community, we believe if any entity wants to use music or books or photographs, they need to get permission, have people affirmatively opt in. So legislatively, I think the copyright law needs to make clear that AI companies can't train on copyrighted works without a license. But there are other things that are being done as well.

The No Fakes Act is probably one of the most talked about AI bills in the U.S. Congress. That's essentially a bill to protect the artists’ name, voice, likeness, image, almost a right of publicity, to protect artists from having their voices ripped off or cloned, and I think there are a series of other transparency factors that are being discussed in the legislative context.

So for instance, there’s having legislation that requires AI companies to keep track of what they use to train their algorithms on, requiring them in some cases to submit new LLMs [Large Language Models] to the copyright office or some other federal agency just to make sure that they're doing things by the book, and transparency laws about labeling and notification. But don't forget that there's a commercial angle to this too, and the industry is reacting very differently to this than they might have done twenty years ago.

How is SoundExchange helping to contain AI’s creative and economic threats to music?

We're doing it in lots of ways. First of all, we represent the entire sound recorded music business in the U.S. I would say even broader than the U.S. Our board represents large record labels, smaller record labels, record label trade associations, both of the artist unions in our country, AFM and SAG-AFTRA, plus artist lawyers, artists themselves, artist managers, and studio producers. So when you look at who we work for, it's pretty much the panoply of everyone involved in the creation of a sound recording.

So we are a powerful voice for the sound recording industry and sound recording policy, and we were one of the founding members of the Human Artistry Campaign. We advocate in alignment for a lot of the industry issues on these AI topics. We talk to the U.S. Copyright Office in support of all the legislation, and we ourselves are trying to educate our constituencies about what's happening with AI and the risks of AI.

One of AI’s greatest dangers is pollution of the royalty stream. If you think about how money flows in the music industry, 85 percent of it is now streaming in the U.S. It's less globally – about 65 percent or 70 percent. 85 or so percent of U.S. recorded music revenue is from streaming, paid out typically from a huge revenue bucket that gets divided among all the streams, and the more AI streams that you throw into that bucket, it starts to pollute the royalty pool. We're at the point now where 100,000 tracks that are AI generated are being uploaded to streaming services every day. Fans and people connect with the humanity of what a musician is and the story behind them and the narrative. It's not just about the music.

Effective communication with the general public is an important part of IP protection. Do you think the public is being properly sensitized to the consequences of AI?

That is a great question. I'll say it's clear that the public is enamored with certain parts of AI. I think the public is sensitized to the dangers of AI. I hope along with that comes the fact that they realize AI runs the risk of really hurting music creators that we represent. I hope that the public recognizes the world will be worse off if we ever got to a place where all we ever listened to was AI music. It would be a repetitive, uneventful place with very new innovations in the music space.

Would you talk about AI’s positive potential for creators?

From an operational standpoint, SoundExchange is a very intensely complicated. operational company, and there are a lot of operational efficiencies that AI can bring to the table in data analytics and fraud detection. There are also opportunities in music production.

But I also think of music resurrection. One of the examples I can use is Randy Travis, who had a stroke [in 2013], hasn't sung for a while, but last year he dropped a new record – again, with consent, credit, and compensation – using a stand-in performer in Nashville. Or Timbaland who always wanted to collaborate with Biggie Smalls and he never had the chance and now he does. Three years ago, a K-pop band [Midnatt] dropped an album in six different languages on the same day. You have situations where someone may have a signature sound and they want to share it with subscribers. Tom Morello, the Rage Against the Machine guitarist, has a very unique sound and he’s worked with an AI company to come up with a licensed product with his permission.

More and more, when you talk to record labels, they believe if the law has the right guardrails, AI has the potential to be a revenue boost for the industry. So there’s a part of me that's intrigued, hopeful, and optimistic on AI. But as I said, before you get to that optimism, we need to make sure that there's a proper foundation that sets guardrails, respects the law, respects the human creator, respects humans give to the creative process, and most importantly, relies on consent, credit, and compensation.