RIAA says labels and fans are aligning against AI music scraping

In short: The head of the RIAA says record labels, artists, and many fans now share the same complaint about AI companies using music without permission, and he is pushing for clearer rules and labels for AI-made tracks.

What's going on

In a new interview on The Verge’s Decoder podcast, RIAA CEO Mitch Glazier said the music industry’s fight has shifted. During the Napster era, the RIAA was widely criticized for suing people who downloaded songs. Glazier says today the bigger conflict is with large AI companies that train their systems on music and other creative work without permission.

Training means feeding huge amounts of material into an AI system so it can learn patterns, a bit like making it read an entire library to learn how to write. Glazier argues that if AI-made music floods services like Spotify and YouTube, it could make it harder for human musicians to earn money and be discovered.

Glazier also discussed “deepfakes,” which are fake audio or video that can copy a person’s face or voice. He said the US does not have a clear federal right that lets people quickly control use of their voice or likeness, and he pointed to the proposed NO FAKES Act as one attempt to create a legal process for takedowns with safeguards.

What to watch

Glazier said labeling will be a key pressure point, so listeners can tell when a track is fully AI-generated. He suggested that labeling only works at scale if major distributors and platforms cooperate, since they are the “gateways” where music gets uploaded and recommended. The next big test will be whether platforms adopt clear, track-level labels and whether courts accept the RIAA’s argument that training AI on music without permission is not fair use.

Source: The Verge AI