The AI music derivative work question has become the question the musician, the platform, the regulator have all been quietly watching, the question the next lawsuit will turn on. The honest framing matters here, because the AI music derivative the platform has been calling the original work sits as the AI music derivative the musician has been quietly recognising as the training data the model has been reworking.
What follows runs as the working version of the field guide. The shorter version is what the working musician, the platform, the lawyer actually have time to read.
What the derivative work question actually is
Here is the working order, by impact. The first runs as the substantial similarity, where the similarity the copyright test asks the court to apply, the similarity between the training data and the AI output, the similarity the court has been measuring through the average observer test the copyright law has been using for decades, the similarity that has been the central question of every AI music lawsuit. The second runs as the transformation, where the transformation the AI performs on the training data, the transformation the copyright law treats as the test for the fair use, the transformation that the AI model performs through the embedding, the attention, the decoding, the transformation the court has been trying to assess without the technical framework the court has been slowly building. The third runs as the market substitution, where the substitution the AI output causes for the training data, the substitution the copyright test has been using since the Sony Betamax case, the substitution the AI music output has been quietly causing for the working musician the AI music has been replacing.
What the 2025-2026 lawsuits showed
Here is the order things have actually landed, in this sequence. The first runs as the training data disclosure, where the disclosure the court has been ordering the model maker to produce, the disclosure that names every training data source the model has been using, the disclosure that has been producing the evidence the rightsholder has been waiting for. The second runs as the settlement pattern, where the pattern the model maker has been following, the pattern of the licensed training data pool the model maker has been negotiating with the major rightsholder, the pool the smaller rightsholder has been waiting to access, the pool the next settlement will likely extend. The third runs as the output level ruling, where the ruling the court has been working toward, the ruling that says when the AI output crosses the line from the inspired by the training data into the substantially similar to the training data, the ruling the model maker has been trying to avoid through the technical safeguard the model maker has been building.
What the working musician should do
Three moves if you are the working musician who has been trying to figure out where the AI music leaves the working musician. Document the training data, where the documentation the working musician should keep, the documentation that shows which track has been used to train which model, the documentation the working musician can produce while the evidence is still accessible, the documentation the lawyer will need when the lawsuit the working musician wants to file lands. Join the collective, where the collective the working musician can join (the Music Publishers Association, the Songwriters Guild, the AFM, the local rightsholder association), the collective the lawsuit the working musician cannot afford to file alone can use, the collective the working musician should be joining now rather than after the next settlement. License the catalogue, where the catalogue the working musician owns, the catalogue the model maker needs, the licence the working musician can offer to the model maker the working musician wants to support, the licence that turns the working musician from the victim into the participant. The musician that documents, joins, and licenses serves as the musician who has positioned the musician for the next round of the lawsuit the industry is working through.
The bottom line
AI music derivative work in 2026 sits as the question the lawsuit will turn on. The substantial similarity, the transformation, the market substitution, those three are what the question is. The training data disclosure, the settlement pattern, the output level ruling, those three are what the lawsuits showed. The document, the collective, the licence, those three are what the musician should do. The musician that does the three is positioned. The musician that has not done the three serves as the musician the next settlement will forget to include.
Sources & Further Reading
All claims in this article are sourced from primary documentation, vendor advisories, and reputable security researchers.
Spotted an error? Email the editor. Corrections are issued with a visible correction note.
Editorial standards. Every article on humanrequired.org is reviewed by a human editor before publication. AI may assist with drafting or research; final editorial control is human. Read the full standards.
