AI is capable of breaking the music industry. But it does not have to.
An op-ed from MusicInfra CEO Björn Lindvall argues that generative AI's most pressing threat to the music industry lies not in creativity but in the sector's ageing infrastructure for tracking royalties and attribution. He warns that the surging volume of AI-generated and AI-assisted music will soon overwhelm systems designed for a simpler era of streams and downloads, unless rightsholders and platforms rebuild them urgently.
Lindvall cites Deezer, which is reportedly receiving around 90,000 AI-generated tracks daily, adding significant strain even though many are demonetised. He argues the bigger challenge is yet to come: as AI-assisted elements blend into human-made songs, and as licensing deals between AI firms and labels require tracking potentially billions of tokens back to rightsholders, the industry will need far more granular attribution than current infrastructure can support. He also points to a widening range of AI-driven apps, games and platforms generating continuous usage data, compounding the reconciliation burden.
- AI's real threat to music is infrastructure, not creativity, argues op-ed
- Deezer reportedly gets ~90,000 AI tracks uploaded daily
- Licensing deals require tracking billions of tokens to rightsholders
New here? Start with this
Music royalties are payments made when music is streamed, downloaded or used elsewhere. They are shared among people and companies with rights in a recording or composition, and depend on records that identify the work and those entitled to payment.
Those records and payment systems were built for a smaller volume of music and simpler ways of listening. AI tools can generate whole tracks or contribute sounds and other elements to music made by people, making it harder to identify who should be credited and paid.
Artists, songwriters, record labels, streaming services and AI companies all have interests in how music is identified and licensed. Accurate tracking matters because it affects whether rights holders receive the payments they are due as music spreads across apps, games and other platforms.
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
The infrastructure case is pragmatic and urgent. Current systems designed for manageable release volumes will collapse under the scale of AI-generated content—90,000 tracks daily to Deezer alone—and future licensing requiring granular token-level tracking demands infrastructure that doesn't exist. Rebuilding this technical foundation is essential, regardless of other policy solutions, or the industry will face reconciliation chaos and unpayable disputes.
The case against
Focusing on infrastructure misses the music industry's core crisis. Musicians face immediate economic threats as AI systems trained without consent displace human work and devalue creativity. The central issues are legal and economic—establishing what AI uses should be permitted and at what compensation—not administrative tracking. Without first resolving copyright, consent, and compensation frameworks, improved infrastructure simply manages inadequate arrangements more efficiently.