In October 2025, the agreement between Universal Music Group (UMG), the largest music major, and UDIO, a generative AI platform specializing in music creation, is much more than a simple technological partnership. It represents a paradigm shift from the conflictual model that prevailed until then between rights holders and AI developers. While recent years have seen a proliferation of lawsuits for alleged infringement (notably cases against Stability AI, Anthropic, or complaints filed by UMG itself against Anthropic), this agreement inaugurates a new era based on the pre-existing and systemic license. It is not just an amicable settlement, but a prospective legal and economic framework that redefines the rules of the game for the exploitation of works in the AI era. This model, which will be deployed through a subscription platform in 2026, could become a reference for the entire music industry, reconciling algorithmic creation and respect for creators’ rights.
For Africa, this is a strategic agreement that could redefine the future of Afrobeat and other so-called “World music” of African origin in the era of musical AI. Indeed, the agreement between Universal Music Group and the start-up Udio represents both a global distribution opportunity and a major legal challenge in terms of intellectual property.
With an estimated annual growth of 18% on streaming platforms, Afrobeat and other African music genres are emerging as the most dynamic genres in the global market. Artists like Burna Boy, Rema, or Ayra Starr accumulate billions of streams, and majors like Universal are heavily investing in African catalogs. In 2023, UMG strengthened its presence in Africa through its subsidiary UMG Africa, with offices in Lagos, Abidjan, and Johannesburg.
In our previous Tribune dated June 23, 2025 – Financial Afrik – When the algorithm sings, the musician is disappointed, we warned that “without clear rules, tomorrow’s silence will no longer be golden: it will be synthetic, dehumanized, and tragically empty.” This agreement outlines a new horizon. Indeed, this innovative device aims to reconcile technological innovation and respect for creators’ rights, while offering users a personalized experience of creation, broadcasting, and sharing music in a legally secure framework.
I. The End of the Counterfeiting Logic: Moving from Feedback to Predictive Feeding
Traditional copyright is essentially retrospective: it intervenes after the fact to sanction unauthorized use. Generative AI, by training on protected data, has placed actors in a legal gray area where the qualification of “illicit copy” for the learning stage was at the heart of the debates.
A. Legalization of Machine Learning (Training)
The Universal-UDIO agreement validates and explicitly legalizes the training process. UMG grants UDIO a license to use its vast catalog (the “training corpus”) to train its AI models. This gesture resolves the thorny issue of the legality of the learning phase, not through an uncertain legal exception (like “fair use” in the United States or the text and data mining exception in Europe), but through contractual and paid consent. It transforms works, previously perceived as simple potentially infringing inputs, into licensable data assets.
B. Creation of a “Right to Train”
Behind the scenes, this act consecrates the emergence of a new implicit economic right for rights holders: the “right to train.” It is not a right provided for by intellectual property law, but it becomes a de facto negotiable prerogative. UMG not only sells music streams; it sells access to its works as the foundation of future creation platforms’ intelligence. This represents a new source of direct and massive valorization for the musical heritage. For example, a new fictional Afrobeat/Reggae track “The year of return” by a possible duo Fela Kuti-Bob Marley, extolling the growing interest of Afro-descendants in Africa, could further perpetuate the influence of these two pioneers among younger generations.
II. Emergence of a New Economic and Governance Model
Beyond the simple license, the agreement establishes an economic and ethical framework that could become the norm.
A. A Novel Redistribution Model
The agreement provides for remuneration for UMG and, by extension, its artists. It is not just a flat fee, but probably a revenue sharing model generated by the UDIO platform. This creates a direct economic causality link between the creation generated by AI and the compensation of the human creators who made this AI possible. It is a concrete response to the historical demand of rights holders to be associated with the revenues of the technologies exploiting their works.
B. Establishment of Collaborative Governance and Transparency Measures
The agreement includes a commitment to develop tools for “traceability” of generated content. Concretely, UMG and UDIO will work together on systems, potentially based on watermarking or digital fingerprinting, to identify if a creation from UDIO has used elements protected by Universal’s catalog. This collaboration is revolutionary: the rights holder is no longer an external policeman but an integrated partner in the design of technological safeguards. This allows for a much more effective proactive protection than post-litigation actions.
III. Prospective Implications: Towards Standardization and New Challenges
The Universal-UDIO agreement is not isolated; it is part of a series of similar partnerships (such as the one between Warner Music and YouTube) and heralds market standardization.
A. The Establishment of a “Two-Speed Model”
This agreement establishes a two-speed future for AI actors:
1. Licensed actors like UDIO, who will have privileged, legal, and quality access to major catalogs, allowing them to develop more efficient and ethical models.
2. Unlicensed actors (independent or “pirates”), who will have to settle for royalty-free or lower quality data, and remain exposed to high legal risks.
This distinction will create a significant entry barrier, consolidating the power of majors and the first AI actors who have signed agreements.
B. Persistent Challenges and Limits of the Model
This new paradigm is not without raising new questions:
• The Digital Divide of Rights Holders: Only the largest catalog holders (the majors) have the power to negotiate such agreements. What about independent labels and self-produced artists? The risk is to create a new form of wealth concentration.
• Status of “Orphan” or Out-of-Agreement Works: AI models will always train on the entire web. The Universal-UDIO license only covers the UMG catalog. The question of using unlicensed works remains open.
• Copyright on AI Outputs: If the agreement settles the input question, the output (the generated music) remains to be clarified. Who is the author? The machine, the user, the platform developer? The agreement could provide specific clauses on the attribution and exploitation of these derivative works.
C. What Producers and Artists of Afrobeat and other African music genres should anticipate
The Universal-Udio agreement could allow Afrobeat to be exported even more widely, integrating AI creation tools accessible to all. But this expansion should not come at the expense of African creators’ rights. It is essential for artists to be involved in negotiations, for their works to be protected, and for economic models to be fair.
1. Preserve the cultural and stylistic value of the genre: Afrobeat is based on complex rhythms, multiple influences (highlife, jazz, hip-hop), and a strong regional identity. Artists must ensure that AI does not dilute this richness by generating stereotyped or decontextualized content.
2. Negotiate clear and remunerative licenses: Producers must ensure that their works used to train AI are covered by transparent contracts, including royalties proportional to usage.
3. Claim neighboring rights on the generated works: If an AI generates a piece based on an Afrobeat artist’s style or voice, they could claim a neighboring right, even if the work is new.
4. Organize collectively: African labels and collectives would benefit from creating collective management entities to negotiate with AI platforms, similar to copyright management societies.
IV. Anticipating Counterfeiting Risks
Despite these safeguards, several legal risks remain:
• Unintentional reproduction: AI could generate works too close to protected originals, leading to infringement disputes.
• Misuse prompting: Users could deliberately encourage AI to reproduce specific styles or voices, bypassing protection mechanisms.
• Uncertainty about the ownership of generated works: The question of whether an AI-generated work is protectable and who it belongs to remains legally open.
V. What Future for Copyright in this New Paradigm?
This model calls for a redesign of copyright:
• Recognition of AI-generated works: The law will need to clarify if and how an AI-generated work can be protected, and under what conditions (human intervention, originality, etc.).
• New neighboring rights: Performers could benefit from neighboring rights on works generated from their voice or style, even if the work is new.
• Hybrid remuneration models: Micro-licensing systems or royalties proportional to usage could emerge to compensate creators whose works have been used to train AI.
• European legislative framework: The AI Regulation (AI Act) and copyright directives will need to integrate these new practices to avoid legal loopholes.
Conclusion: an opportunity to reinvent intellectual property
The Universal-UDIO agreement marks a decisive turning point. It buries the logic of judicial confrontation to inaugurate an era of contractual co-regulation. It transforms copyright from a defensive shield into an offensive lance to capture value in the AI revolution. By legalizing and monetizing the “right to train,” it sets a powerful precedent that will likely set a precedent in future negotiations. However, by institutionalizing a partnership between giants, it lays the groundwork for a new oligopolization of music creation, where access to works from the past will become the key to shaping the sounds of the future. The challenge for public authorities will now be to ensure that this new model, if it becomes widespread, also preserves access and compensation for the entire creation chain, not just its most powerful actors.
The agreement between Universal and Udio is not just a response to a dispute: it paves the way for proactive regulation of creative artificial intelligence. By laying the groundwork for licensed training, it offers an ethical and legal framework for algorithmic music creation. It is now up to legal professionals, legislators, and cultural actors to continue this momentum by reinventing the foundations of copyright to remain a tool for protection, recognition, and valorization of creators in the digital age.
