The music industry is undergoing a seismic transformation as artificial intelligence (AI) reshapes how music is created, distributed, and monetized. Once the subject of heated lawsuits, AI music platforms like Suno and Udio are now in talks with major record labels — Universal Music Group, Sony Music Entertainment, and Warner Music Group — to negotiate licensing deals that could legitimize AI-generated music. This shift marks a pivotal moment, moving the conversation from whether AI music will be regulated to how it will be integrated into the industry. Yet, with these negotiations happening behind closed doors, critical questions arise: Who will benefit from this new era, and who will be left behind?

From Courtrooms to Contracts: A New Chapter

In mid-2024, the Recording Industry Association of America (RIAA) filed lawsuits against Suno and Udio, alleging “massive” copyright infringement for training their AI models on copyrighted songs without permission [1]. The labels claimed these platforms produced music that mimicked protected works, threatening artists’ livelihoods. Suno and Udio countered that their use of copyrighted material fell under fair use, arguing their outputs were transformative [6, 9]. By early 2025, however, the tone had changed. Reports from Billboard and Bloomberg indicate that the majors are now negotiating licensing agreements with these startups, potentially involving fees, royalties, or even equity stakes [3, 4].

This pivot recalls the early days of music streaming. When Spotify launched, major labels secured licensing deals that included advance payments, minimum guarantees, and equity stakes, giving them leverage over platform policies and a share in its growth [7]. Those deals, however, often sidelined independent artists, songwriters, and smaller publishers, who had little say in how revenue was distributed. Today’s AI music negotiations risk a similar outcome. With the “Big Three” labels leading the charge, smaller stakeholders may be forced to accept terms they didn’t help shape, perpetuating industry inequities.

The Attribution Challenge: Who Gets Paid?

Unlike streaming platforms, which distribute existing songs, AI platforms like Suno and Udio generate new music that competes directly with human-created works [8]. This raises a complex issue: how do you license something that’s inspired by, but not identical to, copyrighted material? Fair compensation hinges on attribution — tracing AI outputs back to the training data — but current technology makes this nearly impossible. Generative AI models use probabilistic algorithms to create music based on patterns, not direct copies, leaving no clear “digital fingerprint” to identify which songs or artists influenced a track [2].

Without attribution, splitting revenue among rights holders becomes a logistical quagmire. Should licenses involve flat fees for access to training data, or royalties based on generated outputs? If there’s no link to the original works, how do you determine who gets paid? Clean metadata — accurate links between songs, recordings, and songwriters — will be critical for those hoping to claim revenue, but only well-resourced players, like major labels, typically maintain such data. Smaller publishers and independent artists, often lacking robust metadata systems, could be disadvantaged, much like they were during the streaming era’s digitization push [10].

The absence of attribution technology also complicates royalty distribution. In streaming, performance data drives payouts, but AI-generated music lacks clear usage metrics. A flat-fee model might favor labels with large catalogs, while a per-track royalty system could undervalue contributions from lesser-known artists. Until attribution becomes feasible, the industry risks creating a system where only the most powerful players thrive.

Artist Agency: A Fading Hope?

A pressing concern is whether artists will have control over how their work is used in AI training. Many AI models, including those of Suno and Udio, have already been trained on copyrighted music, often in jurisdictions with lax copyright laws to avoid legal scrutiny [5]. This “forgiveness over permission” approach means the data is already ingested, making retroactive opt-outs challenging. Retraining models to exclude opted-out works is costly and technically complex, leaving artists with limited leverage [2].

Even if opt-out options are introduced, they may apply only to future training, not existing datasets. This mirrors challenges in other AI sectors, where companies like OpenAI have faced criticism for using copyrighted material without consent. Artists and advocacy groups, such as the Music Workers Alliance, are demanding consent, credit, and compensation, but their influence pales compared to major labels [8]. Without coordinated pressure from lawmakers or collection societies, the emerging AI music framework may prioritize corporate interests over creators’ rights.

An Ethical Path Forward: Consent-Based Models

Amid these challenges, some platforms are exploring a more ethical approach. For example, Delphos.ai trains its AI only on music explicitly provided by users with clear consent, ensuring attribution and legal usability [7]. This model allows rights holders to choose whether their work is included, share in revenue, and avoid copyright disputes. While slower to scale than web-scraping models, consent-based platforms could set a standard for sustainability, especially as regulators tighten oversight. Forbes notes that such models empower creators by giving them agency, contrasting with the opaque practices of some AI startups [7].

Consent-based systems also address a looming question: is AI-generated music even copyrightable? Works created without human input may not qualify for copyright protection, creating legal risks for platforms and users [2]. By involving rights holders from the outset, ethical models mitigate these uncertainties, offering a blueprint for a creator-friendly AI music ecosystem.

The Future of AI in Music: Innovation or Disruption?

AI music is no longer a futuristic concept — it’s reshaping the industry now. The future will likely feature a mix of licensed platforms, ethical innovators, and market disruptions:

• Licensed Ecosystems: Deals between majors and platforms like Suno and Udio could create a regulated AI music market, with licensed outputs competing alongside human works. Advances in attribution technology, such as AI fingerprinting or blockchain-based metadata, could enable fairer revenue splits, though these are years away [8].
• Ethical Platforms: Consent-based models like Delphos.ai may gain traction, especially if consumers and regulators demand transparency. These platforms could carve out a niche for high-value, legally clear AI music, appealing to creators and businesses alike [7].
• Market Saturation: Without regulation, AI-generated music could flood platforms, lowering production costs but threatening human artists’ income. MIT Technology Review warns that unchecked AI could “saturate the market with machine-generated content,” disrupting traditional revenue streams [8].

The balance between innovation and fairness will define AI music’s trajectory. If attribution remains elusive, major labels with legal and metadata resources will dominate, while independents struggle. However, technological breakthroughs or regulatory mandates could level the playing field, giving smaller players a chance to compete.

AI Legislation: A Patchwork Approach

AI legislation is struggling to keep pace with the technology. In the U.S., the Copyright Office’s 2024 report stated that training AI on copyrighted works without permission likely violates fair use, supporting the labels’ lawsuits against Suno and Udio [2]. However, the recent dismissal of Copyright Office Director Shira Perlmutter has sparked fears that a new, AI-friendly director could soften copyright protections under a Trump administration [7]. This uncertainty complicates the legal landscape for AI music.

Globally, copyright enforcement varies. Some AI companies train models in countries with weaker regulations, creating cross-border challenges reminiscent of music piracy [5]. In March 2024, Tennessee passed the ELVIS Act, the first U.S. state law protecting musicians from unauthorized AI use, such as voice replication [8]. Other states may follow, but a lack of federal standards leaves the industry in limbo. In the European Union, the AI Act, effective August 2024, requires transparency from AI developers, but its music-specific implications are unclear [7].

Without global coordination, licensing deals will set the industry’s rules, potentially favoring major labels over creators. Advocacy groups are pushing for stronger protections, but their impact is limited compared to corporate lobbying. The outcome of these legislative efforts will shape whether AI music becomes a tool for creativity or a source of exploitation.

Building a Fair AI Music Future

The shift from lawsuits to licensing signals a new era for AI music, but the stakes are high. Will all rights holders be compensated fairly? Will artists have a say in how their work is used? Will transparency be a cornerstone of the system, or an afterthought? The industry must address these questions to avoid repeating the streaming era’s mistakes, where a few reaped massive rewards while others were left behind.

To create a fair AI music ecosystem, stakeholders should prioritize:

• Attribution Technology: Invest in systems to trace AI outputs to training data, ensuring equitable revenue distribution.
• Creator Control: Implement opt-in or opt-out mechanisms, with clear policies for existing datasets.
• Inclusive Frameworks: Involve independents, songwriters, and artists in licensing negotiations to balance power dynamics.

AI music has the potential to democratize creativity, but only if built on a foundation of fairness. As the industry negotiates its future, the choices made today will determine whether AI becomes a partner to artists or a rival.

Citations

[1] Recording Industry Association of America, June 24, 2024.
[2] WIRED, June 24, 2024.
[3] Billboard, June 2, 2025.
[4] Bloomberg, June 1, 2025.
[5] Associated Press, June 24, 2024.
[6] Music Business Worldwide, August 5, 2024.
[7] Forbes, June 6, 2025.
[8] MIT Technology Review, June 27, 2024.
[9] Reuters, August 1, 2024.
[10] Billboard, June 3, 2025.