The Music Industry's AI Infrastructure: What Blues Artists Should Know About Training Data

The legal battle unfolding between major record labels and AI music startups is one of the most consequential disputes in the history of the music industry. For blues artists, a community whose musical catalog spans everything from pre-copyright public domain recordings to contemporary protected works, the implications are both urgent and specific.

In May 2026, Udio, the AI music startup, filed a formal response to the Sony Music copyright infringement lawsuit that had been working through the courts since 2024. The response included a direct admission: "Udio admits that it obtained audio data from YouTube for use as training data," and specified that "it acquired some of its training data by utilizing YT-DLP", a tool for downloading YouTube content.

For blues artists and the labels, publishers, and estates that hold blues catalogs, this admission crystallizes what was previously suspected: AI music systems have been trained on vast quantities of audio, including music that was available on YouTube whether through authorized uploads or not.

What Udio and Suno Admitted

The legal history is important to understand. In August 2024, both Udio and Suno, the two most prominent AI music generation platforms at the time, effectively admitted in court filings that they had trained their models on copyrighted music. Music Business Worldwide reported that Suno described its training data as "essentially all music files of reasonable quality that are accessible on the open Internet."

Both companies argued that this use falls under "fair use" provisions of US copyright law, specifically that their models create "intermediate" copies that are never heard by users, and that the outputs of their systems are non-infringing new creations.

The RIAA called this "a major concession of facts they spent months trying to hide," noting that the companies had acknowledged "massive unlicensed copying of artists' recordings."

The YouTube scraping allegations against Udio were added by the labels in an amended complaint in September 2025. Udio responded in May 2026 by admitting the YouTube audio use while arguing that US copyright law does not prohibit downloading publicly available videos, an argument framed around fair use doctrine.

Why Blues Catalogs Are Particularly Exposed

The blues raises copyright questions that are more complex than almost any other American genre. The relevant issues involve multiple layers:

Pre-1927 public domain recordings. The music industry has operated under the assumption that pre-1927 sound recordings entered the public domain under the Music Modernization Act of 2018. Many foundational blues recordings fall within or near this window. While these recordings are legally available for use, the lines are not always clean, especially for recordings that exist in multiple versions, some of which are public domain and some of which are not.

The 1950s, 1970s electric blues catalog. The commercial blues recordings from this period, the Chess Records era, the Stax and Volt sessions, the early recordings of artists who defined electric blues, are heavily protected under copyright. Many of these recordings are available on YouTube through authorized label uploads, which may mean they are among the "publicly available sources" that Udio described in its court filings.

Blues as AI training "raw material." Blues is a particularly rich genre for AI training because of its structural regularity, 12-bar progressions, established melodic patterns, recognizable scales, combined with the expressive variation that makes each performance distinct. An AI trained on blues recordings would develop a sophisticated model of the genre's harmonic and rhythmic grammar. This is valuable training data, and it is why the genre's catalog is a target.

Independent and estate-held blues catalogs. Unlike major label catalogs, many blues recordings are held by small independent labels, heirs, or estates that lack the legal resources to pursue individual copyright claims. Class action lawsuits, like the one filed by country musician Tony Justice against both Suno and Udio in June 2025, reported by Music Business Worldwide, are designed to aggregate the claims of independent artists who cannot pursue individual actions.

The Copyright Office's Position

The US Copyright Office issued a significant guidance document in May 2025: Copyright and Artificial Intelligence, Part 3: Generative AI Training. According to the class action lawsuit filed by Tony Justice, as reported by Music Business Worldwide, the report "emphasized that the fair use doctrine does not excuse unauthorized training on expressive works (e.g., music) particularly when those works are used to generate substitutional outputs that may replace the originals in the relevant marketplace."

This is not a final legal ruling, courts are not bound by Copyright Office guidance, but it represents the most authoritative administrative position on the question, and it is unfavorable to the AI companies' fair use argument.

The core tension is straightforward: if an AI model is trained on blues recordings and can then generate new blues-sounding music that competes in the same marketplace as the original recordings, the "fair use" defense becomes significantly harder to sustain. The Copyright Office's analysis supports this view.

What Blues Artists Can Do

The legal landscape will continue to evolve for years as the Udio and Suno cases work through the courts. But blues artists and their representatives do not need to wait for final resolution to take protective steps.

Register your copyrights. Copyright exists from the moment of creation, but registration provides the legal framework necessary to pursue infringement claims and to participate in class action lawsuits. The US Copyright Office registration process is accessible at copyright.gov and is the foundational step for any artist who wants legal standing in AI-related disputes.

Monitor your catalog on YouTube. YouTube's Content ID system is the primary mechanism for detecting unauthorized use of your recordings on the platform. Ensure that your distributor or rights management company has registered your catalog with Content ID. If your music is on YouTube without authorization, through uploads by fans, archivists, or AI companies scraping the platform, Content ID can identify and manage these instances.

Understand what's in the public domain. If you hold rights to recordings that may be approaching or have crossed into the public domain, consult with a music attorney to understand the implications. What is public domain for one use may not be public domain for another.

Stay engaged with the class action process. The Justice v. Suno/Udio class action is designed to include independent artists "whose rights have been trampled the most," as Tony Justice's filing stated. If you believe your recordings were used as AI training data, consulting with an attorney about participation in class action or individual legal action is appropriate.

Document your creative process. Maintaining records of your recording sessions, composition process, and release history is valuable in any copyright dispute. For blues artists who have recorded extensively, organizing this documentation now, before disputes arise, reduces friction significantly.

The Broader Implications for Blues Culture

The AI training data dispute is ultimately about what music creation means in an era where algorithms can generate stylistically convincing content at scale. Blues, as a genre with deep oral and improvisational traditions, has always transmitted knowledge through performance and imitation, but always with the acknowledgment that a living human being was making choices, drawing on experience, and expressing something specific.

AI-generated blues cannot replicate this. It can simulate harmonic patterns and rhythmic feels with increasing technical accuracy, but the music it produces lacks the biographical and emotional specificity that makes blues meaningful. This is not merely a philosophical point, it is a commercial one. As the Copyright Office noted, AI outputs that compete directly with original recordings in the same marketplace raise serious fair use questions.

At Mollohan Production Inc., the implications of AI training data disputes are part of the ongoing conversation with artists about catalog protection and rights management. For blues artists with existing recordings, whether recently produced or part of a longer career, understanding what protections exist and how to use them is increasingly essential to protecting the economic value of the music. The question is not whether this issue matters to blues artists. It does. The question is how prepared each artist is to navigate it.

---

FAQ

What did Udio admit about its AI training data? In its May 2026 court response, Udio admitted: "Udio admits that it obtained audio data from YouTube for use as training data," and specified that "it acquired some of its training data by utilizing YT-DLP." This was part of Udio's response to Sony Music Entertainment's copyright infringement lawsuit. (Music Business Worldwide)

What is the fair use defense that Udio and Suno are using? Both companies argue that their use of copyrighted recordings to train AI models constitutes "fair use" under US copyright law, specifically that the models create "intermediate" copies that are never heard by users, and that the AI outputs are non-infringing new creations. The US Copyright Office's 2025 guidance expressed skepticism about this argument when AI outputs compete directly with original recordings in the marketplace.

Why are blues catalogs particularly at risk from AI training data issues? Blues has an unusually complex copyright situation: some foundational recordings are in the public domain, while the commercially significant 1950s, 1970s electric blues catalog is heavily protected. Blues is also structurally attractive as AI training data because of its regular harmonic patterns. Additionally, many blues rights are held by small independents and estates that lack the resources to pursue individual copyright claims.

What is the Copyright Office's position on AI training data and fair use? The US Copyright Office's May 2025 report, Copyright and Artificial Intelligence, Part 3: Generative AI Training, stated that the fair use doctrine does not excuse unauthorized training on expressive works "particularly when those works are used to generate substitutional outputs that may replace the originals in the relevant marketplace." This guidance is not legally binding but represents the most authoritative administrative position on the issue.

What can independent blues artists do to protect their recordings? Register copyrights with the US Copyright Office, ensure catalog registration with YouTube's Content ID system through a distributor or rights management company, understand which recordings may be approaching public domain status, stay informed about class action litigation against AI companies, and consult a music attorney about individual legal options.

---