Major record labels that only two years ago sued generative-AI music companies for allegedly training on copyrighted recordings are increasingly striking licensing agreements with those same platforms, transforming a copyright confrontation into a potentially lucrative new business model. Suno has launched models using licensed music from Warner Music Group, BMG and Believe, while Udio has agreements involving Universal Music Group, Warner and other rights holders; additional arrangements are bringing licensed AI remixing and music creation to other platforms. The industry says participating artists can opt in, receive credit and share in new revenue, but crucial financial details remain largely undisclosed, including compensation formulas and the degree to which individual performers can actually refuse participation when labels control their recordings. Meanwhile, musicians, songwriters and unions are challenging whether labels possess all the rights necessary to authorize AI training and reuse, particularly where session performances, songwriting copyrights, voices, likenesses and publicity rights are concerned. The central question is therefore shifting from whether AI will enter the music business to who controls its entry—and whether the people who created the underlying music will receive a meaningful share of the value generated from it.
Key Takeaways
- Major music companies are moving from litigation toward licensing, with Suno, Udio and other AI platforms gaining authorized access to portions of commercially valuable music catalogs even as separate copyright disputes continue.
- Labels and AI companies promise artist consent, attribution and compensation, but payment formulas generally remain confidential, while existing recording and publishing contracts may give rights holders considerably more control over AI licensing decisions than individual musicians possess.
- The emerging battle is increasingly about property rights and compensation: musicians and songwriters contend that label ownership of recordings does not necessarily confer unlimited authority over performers’ services, compositions, voices, identities or contractual rights when those assets are repurposed for generative AI.
In-Depth
The music industry’s embrace of generative AI is becoming less a technological experiment than a test of who actually owns the economic value embedded in recorded music. Major labels once portrayed unauthorized AI training as a threat to copyright and human creativity. Now several are negotiating licenses with AI companies, betting that controlled access can turn disruption into another revenue stream.
That approach has obvious commercial logic. If generative music cannot realistically be eliminated, rights holders have an incentive to license it, establish contractual safeguards and collect revenue rather than allow technology companies to appropriate valuable catalogs without payment. Recent agreements increasingly promise licensed training, artist participation, attribution and compensation. Some also impose safeguards such as watermarking, download restrictions and controlled distribution.
The unresolved issue is whether the interests of labels and creators are actually identical. They are not necessarily so. A record company may control a master recording while songwriters or publishers control the composition. Session musicians may have contractual reuse rights, while performers can possess separate interests involving their names, voices and likenesses. The American Federation of Musicians has already sued major labels, arguing that AI licensing triggered contractual obligations to compensate musicians for new uses.
That distinction matters. Responsible AI licensing could create a legitimate new marketplace, but only if property rights remain meaningful throughout the chain. If labels and technology companies capture most of the financial upside while the people whose performances made the catalogs valuable receive little compensation or practical control, licensing will have solved the industry’s copyright problem without necessarily solving the creators’ problem.
Sources
- https://www.afm.org/2026/08/summary-of-the-afm-v-universal-warner-and-atlantic-new-use-litigation/
- https://www.musicbusinessworldwide.com/suno-inks-global-licensing-deal-with-believe/
- https://www.universalmusic.com/universal-music-group-and-elevenlabs-announce-multi-year-strategic-agreement-beginning-with-a-new-licensed-ai-music-creation-platform/
- https://www.universalmusic.com/spotify-and-universal-music-group-announce-landmark-licensing-agreements-for-fan-made-covers-and-remixes/