CISAC study: generative AI risks costing music artists ~20% of their revenues by 2028; Abba's Björn Ulvaeus warns that AI “has the power to cause great damage”
‘Very unfair’ for tech companies to use artists' work without paying them, says pop legend and head of advocacy group
Context & Ripple Effects
The surrounding coverage traces a long-running dispute over who owns, controls and is paid for machine-made or machine-assisted music: earlier reporting flagged unresolved authorship and compensation questions, while the RIAA’s action against generative-music startups brought training-data use into a more adversarial phase.
CISAC’s estimate gives that rights debate an explicit economic frame. It matters because later coverage shows labels pursuing AI licensing arrangements modeled on user-generated-content revenue splits, suggesting that parts of the industry are trying to convert a conflict over unlicensed use into a commercial rights market.
First-order effects
- The study arms CISAC, Ulvaeus and artist advocates with a concrete projected revenue exposure as they press technology companies to pay for use of artists’ work.
- Artists and rightsholders face a sharper distributional risk: generative music can substitute for or compete with catalog and new releases without a settled compensation mechanism.
Second-order effects
- Labels, publishers and music platforms gain more incentive to distinguish licensed from unlicensed AI music, both to protect royalty pools and to avoid amplifying content whose inputs lack permission.
- Generative-music developers face greater pressure to seek licenses or demonstrate provenance; the alternative is a more contentious market shaped by litigation and platform moderation demands.
Third-order effects
- If licensing becomes the prevailing response, music AI may develop around negotiated access to catalogs and revenue-sharing rather than unrestricted training on available recordings.
- The durable fault line will be bargaining power: well-organized rightsholders may secure terms first, while independent creators’ ability to obtain attribution, payment and enforcement remains less certain.
The trend: Generative AI is pushing music from a copyright debate toward a market for licensed training data, provenance and royalty-sharing.