Hold up — I had a great time talking to the RIAA guy

Mitch Glazier, chairman and CEO of the Recording Industry Association of America, sat down with Nilay Patel on the Decoder podcast for a conversation published in early 2026. The framing is unusual: both men began their careers on opposite sides of the Napster wars, with Patel defending students sued over file sharing and Glazier, then a young lawyer at the RIAA, working on the industry's enforcement campaign. Two decades later they largely agree on the problem, which Glazier describes as big AI companies training on everyone's work without permission and filling feeds with machine-made content.
Glazier's account of the Napster era is the most concrete part of the discussion. He says the trade group ran a settlement program under which students identified by their universities' IP addresses were told they would face district court in Chicago and should settle for $5,000, a figure he says was set to break even on costs and deter copying. He recalls being sent to college auditoriums to persuade students against peer-to-peer software, a hostile assignment he once opened by walking onstage with a garbage bag over his head after following Chuck D, who was there promoting peer-to-peer. He describes those campus arguments as among the best conversations he has had about technological change, even though the period was heated.
On the current fight, Glazier's argument rests on a simple theory of why copyright exists: to create more opportunities for more art in the future by letting creators earn a living. He contends that machine-generated music, which he says can be produced in seconds without lived experience and, in his view, without a soul, risks diluting or drowning out the incentive for humans to keep making art. He casts the present moment as different from Napster because labels, artists and fans are aligned rather than at odds, and he characterizes the industry as fighting to preserve the human connection between artist and fan rather than merely defending a business model.
The conversation ranges across the areas where the RIAA is both lobbying and litigating: deepfakes and the free-expression complications of banning them, and the practical difficulty of detection and labeling systems if major platforms such as Spotify and YouTube do not participate. The exchange is a podcast interview rather than an announcement, so there are no new policy proposals, products, filings or numbers beyond the historical $5,000 settlement figure and the organizations and platforms named.
Patel supplies the economic frame. He notes that in the CD era value was tied to a scarce physical object, that digital distribution drove the cost of copying a song to near zero and forced the subscription streaming model into existence, and that AI pushes the cost of creation itself toward zero. That produces a glut of supply, and the open question he puts to Glazier is where scarcity returns: through lobbying and law, through platform labeling, or through fans rejecting machine-made work. Patel also observes that only the biggest artists and the platforms seem to be making real money from art today, and that the middle class of artists may be hollowed out.
Glazier's answer is that fans will always win, because any service, AI or streaming, needs people to pay, usually by subscription, and they will not pay if they feel they are losing something. He cites focus groups and polling indicating that listeners want a human connection to artists, and argues that the persona and brand around a performer matter as much as the recordings. He distinguishes fully machine-made art, where he expects that preference to persist, from assistive AI, which he suggests may come to be seen the way synthesizers or Auto-Tune are now.
This suggests the RIAA's near-term leverage lies less in technical blocking and more in disclosure and market pressure. If labeling is uneven across large platforms, listeners cannot reliably exercise the preference Glazier describes, and detection of AI-generated audio remains an unsolved technical problem even for well-resourced platforms. His bet on consumer sentiment also puts weight on a market signal that is hard to measure and easy to game, since a track assembled with AI assistance sits in a gray zone that the synthesizer comparison deliberately blurs.
Why it matters: teams building generative audio, music recommendation or content labeling should expect continued RIAA litigation and lobbying pressure, and should treat provenance and disclosure as product requirements rather than optional features, since the industry's stated strategy depends on listeners knowing whether a track is human. Developers on the creative-tools side face a less certain path: Glazier's own framing suggests assisted creation may be tolerated where fully synthetic output is not, an inference from his comments rather than a stated safe harbor. Platforms that stay out of any labeling regime risk becoming the target that makes the rest of the system moot.
Based on reporting from the original publisher. Visit the source for full context and later updates.
Publisher excerpt
Today, I’m talking with Mitch Glazier, who is CEO and chairman of the RIAA. This conversation was a real full-circle moment for me: I was in college when Napster hit, and the RIAA became everyone’s villain for suing kids for downloading music. That was when I decided to go be a copyright lawyer, which set […]