Between Two Classes: Mitch Glazier

“The machines handle the math, and the humans handle the magic.”

This Between Two Classes editor came out of retirement for the distinct pleasure of interviewing Mitch Glazier, Chairman and CEO of the Recording Industry Association of America (RIAA). Mitch leads the trade organization representing the three major recorded music companies—and, by extension, an $11 billion American industry that shapes culture in the U.S. and abroad. Before stepping into the top role, Mitch spent two decades at the RIAA shaping music policy; before that, he served as Chief Counsel to the House subcommittee overseeing intellectual property law. On April 9, Mitch visited HBS for a talk to students from HBS, Harvard College, and Berklee, followed by a student dinner. Our conversation spans the RIAA’s ongoing lawsuit against Suno, the “authenticity backlash” Mitch sees coming, why investing in a musician is fundamentally different from greenlighting a film, and why—despite all the noise around AI—fans still give him hope.

Let’s start at the top. Looking at the AI landscape today, what do you see as the most urgent policy issue facing the music industry?

Glazier: There are really four buckets. The first—and the biggest one—is the training of AI models. This is the Suno issue. This is why we’re in litigation against Suno. Our argument is that Suno scraped the internet, made reproductions of recorded music owned by the companies we represent, and didn’t ask permission. You’re not supposed to have the right to make a copy if you don’t have permission. That’s why it’s called copyright.

Their argument back is fair use—that they’ve transformed the input into something new. But the fair use analysis includes a factor the courts consider most important: you cannot make a copy that has a negative impact on the potential market for the underlying work. When you upload hundreds of thousands of AI tracks onto Spotify, which pays artists out of a single pooled revenue stream, and those tracks get incorporated into the recommendation algorithm, by definition you are reducing the pool for every human artist whose work was used to train the model.

In the Napster era, public opinion was something like 70-30 against the industry. Today, it seems closer to 90-10 in your favor. What flipped?

Glazier: AI is fundamentally different from previous technologies so we’re sorting out how to deal with it even as the technology rapidly moves—it isn’t just disrupting the music industry, it’s disrupting every industry. And it isn’t just affecting people on the factory lines like other automation has. It’s affecting white-collar workers–lawyers, politicians, people whose job it is to create guardrails and protect labor.

People don’t want to put the genie back in the bottle, but they also realize that only this much of the genie is out. Maybe it’s time to actually try to prevent the bad things from happening before the whole thing is out. And don’t underestimate how important art is to people. There are two things you talk about at the kitchen table: politics and entertainment. When you take what people view as part of their identity and you threaten it, they react.

Part of the argument for moving fast on AI is the race with China. How do you respond to that framing?

Glazier: We’ve analyzed a lot of data, because it’s the same question a lot of people in the administration are asking. Do you want to win in the short term or in the long term? If you look at the data going back, countries that have continually protected intellectual property are the only ones who have won in the long run. Any country that has granted broad exemptions to property rights for short-term gain has lost out. These machines need food. If people stop creating the food, the machines can’t eat. You end up with synthetic content training on synthetic content, and that leads to model collapse.

There’s also a trade argument. American music is by far the largest music export in the world—we’re a bigger music exporter than the next six biggest exporters combined. If we create an exception to copyright, China will follow suit. The difference is we won’t be sucking up any Chinese music, but they’ll be sucking up all American music. Singapore has already done a version of this. Chile has tried. We’re working very hard to prevent it from catching fire elsewhere.

You’ve described artist engagement as your “super weapon.” Can you talk about how the Elvis Act in Tennessee grew into the federal No Fakes Act?

Glazier: Weirdly, in the U.S. you don’t have a federal right to yourself. You don’t have a right to your own voice. You don’t have a right to your own image. Most state laws around name, image, and likeness are geared toward commercial purposes—you can’t put Elvis on a T-shirt, you can’t put Michael Jordan on a Wheaties box—but the general construct is that anybody can take anybody else’s voice for a non-commercial purpose and do whatever they want.

We went to Tennessee, which happens to like musicians very much, and passed the Elvis Act—the first state law to extend that right of publicity into a generalized right to your own voice and likeness, tailored to AI. That made the tech companies very nervous, because suddenly they had to deal with fifty different state laws, and it got them to the table in Washington. Two years later, we’ve locked in the negotiated text of the No Fakes Act, which will create the first federal IP right in your voice and likeness. Artist engagement is what got us there. Natalie Grant testified in front of a Tennessee state representative who was against us. She said, “I believe I was created in the image of God. So if you steal my image….” We got his vote.

Film studios seem to have a much bigger appetite for using AI to cut costs, even shaving $50 million off a $200 million VFX budget. Is music headed the same way?

Glazier: Film and music are fundamentally different industries. In film, you don’t invest in a person—you invest in a film. Every film is its own corporation. The director delivers the final product. In music, you’re partnering with and investing in a person. The artist is the director and the actor. You invest in them because you like their original creativity, and your job is to help them make commercial and creative decisions while keeping them authentic—because authenticity is the reason you invested in them in the first place.

There’s a guy in Hollywood who does massive audience testing for new releases. When I asked him about AI, he said, “You either connect with an audience, or you don’t.” When you efficiency-gain your way to a movie that has no resonance, you don’t know it until it premieres. 

You’ve talked about an “authenticity backlash” coming. How does that interact with how labels should be making bets today?

Glazier: There will be a waterfall of AI-generated slop on streaming services, because it’s efficient and because there’s always a gold rush toward anything new. But fans will seek out authenticity. What you can monetize forever now is the catalog. In the old days, the term was two weeks of sales at Tower Records, followed by return CDs ending up in a warehouse. Now streaming makes the catalog evergreen—you can earn money on Bruce Springsteen thirty or forty years from now. That’s why Goldman Sachs and the venture capital firms are getting into music.

Would you rather have pennies forever, or five dollars only today? From an investor’s perspective, it’s pennies forever. That changes the calculation for labels, for artists, and for what kind of music gets made.

Last one. What gives you optimism?

Glazier: Fans. I truly trust fans. For all the stuff we do, all the games we play, all the strategies we come up with—at the end of the day, I think there’s something magic that happens between an artist and their fans. If something breaks through on a long-term basis, it’s because of a human connection.

I don’t think AI is a novelty. It’s here to stay, and it will be used in a lot of ways. But if a fan doesn’t feel there’s an authentic connection between the person who is the artist and the music coming out, I just don’t think it’s a long-term viable money maker—not for the artist, not for the industry, not for anybody. So, to take a quote from the Hypebot poll of musicians: “the machines handle the math, and the humans handle the magic.”

Peter Sykes (MBA ’26) graduated from Georgetown University in 2015 with a degree in Economics. Prior to HBS, he worked in strategic operations at Reservoir Media, a publicly traded indie music company. Prior to that, Peter held positions at Morgan Stanley, working in both capital markets and at the Morgan Stanley Children's Hospital.