In an earlier article, I posed nine questions to music industry leadership about AI and music. The fourth question was: What are the future economics of creative work?
If an AI-generated output is influenced by many works in a training corpus, how should value be allocated? If a musician is largely responsible for a creation but uses AI tools, what portion of the creation is rightly theirs? These are questions about compensation but beneath them sits a more difficult discussion: attribution.
Some companies argue that detailed attribution is unrealistic because AI training and use do not operate by copying a sample from a single identifiable source. Others argue that probabilistic attribution, similarity analysis, model logging, retrieval records, and metadata can still create useful signals.
The right standard may not be able to reflect absolute certainty. It need only be fit-for-purpose.
From Buyouts to Participation
A one-time buyout model cannot survive as the default future for AI music. Instead, participation in revenue must grow as services grow, be revisited as use cases change, and recognize that catalogs, performances, compositions, metadata, and artist identities may contribute differently to overall product value.
The economic question is therefore not only how much AI companies should pay. It is how payment can be tied to contribution in a system where influence is distributed across many works.
In many music markets, from collective licensing to performance royalties, payment systems already offer examples of imperfect but workable allocation systems. Markets operate using samples, proxies, reporting thresholds, and negotiated rules. The question is whether AI attribution can become reliable enough to support trust and market participation.
Published economic studies underscore the urgency. CISAC and PMP Strategy projected that 24 percent of music creators' revenues could be at risk by 2028 if generative AI substitution grows without adequate safeguards and participation mechanisms.[1] GEMA and SACEM's creator study similarly identified demands for transparency, consent, and a fair share of AI revenues.[2]
These studies should not be treated as definitive forecasts, but they make clear that compensation design is not a technical side issue. It is central to whether the ecosystem remains something worthwhile that human creators should continue to invest in.
More Than a Payment Conversation
Attribution is more than a money problem. It is a trust problem, a transparency problem, and a cultural memory problem.
Musicians want to know whether their work has been used, and for what purposes. Listeners may want to know whether a voice is real or whether the recording they are listening to is the genuine article and not an AI clone. Rights holders need records to audit payments. Educators and historians need to understand how musical materials circulate. If AI tools weaken the chain of credit, they will damage not only income but the social and cultural meaning of musical authorship and trust in the industry.
The Answers Are Being Built
The music ecosystem infrastructure already contains partial answers. DDEX's Recording Information Notification standard was designed to communicate metadata about studio-session entities and contributors from creation environments into the music value chain.[3] The Coalition for Content Provenance and Authenticity (C2PA) provides an open technical standard for digital content provenance and edits.[4]
Neither standard solves AI attribution by itself, but both point in the right direction: information should be captured as early as possible and carried through interoperable systems, rather than reconstructed after a work has already been released. The standards are absolutely necessary, but on their own they are not sufficient for a trusted ecosystem.
Compensation and Attribution Solutions Require Collaboration
For the Music Technology Coalition, compensation and attribution should be treated as design challenges, not only as legal disputes.
The Coalition could test attribution methods, compare metadata schemas, prototype creator dashboards, study revenue-pool models, and convene publishers, labels, PROs, distributors, AI companies, and DSPs around shared reporting fields. It could also examine equity participation and non-royalty value. For instance, if catalogs and creative labor increase the enterprise value of AI companies, participation may need to extend beyond per-output royalties.
The central principle is simple: creative work should not disappear into the machine. Even where technical attribution remains uncertain, the ecosystem can build better records, better disclosures, and better payment proxies.
Transparency will not automatically create fairness, but fairness is impossible without it.
Do you want to learn more about the challenges facing the music technology ecosystem? Then head over to Unity Gain – our hub dedicated to the music tech industry, with insights on issues from the increasingly important role of AI in music to issues around interoperability, transparency, and trust.
[1] CISAC and PMP Strategy (2024), Global economic study shows human creators' future at risk from generative AI.
[2] GEMA, SACEM and Goldmedia (2024). AI and music: Generative artificial intelligence in the music sector.
[3] DDEX, Recording Information Notification (RIN).
[4] Coalition for Content Provenance and Authenticity, C2PA: Verifying media content sources.