It feels like discussions about how to effectively use AI are everywhere. But what would you do if you found out AI was giving your published work to others for free? That’s exactly the concern raised by a group of Chicago artists who claim that AI platform Mureka used and gave away their music without permission. Now, the artists are suing the AI company for copyright infringement. The question is, can AI companies assert a fair use defense to get out of liability when they infringe on copyright? Courts should say no.
Introduction
You might be most familiar with the concept of copyright as a legal right that protects written works if it meets certain criteria, such as being original, creative, and fixed in a tangible form of expression. However, copyright protection is not stuck in the 2D world. It extends to cover a wide range of creative works, like sculptures and even music.
AI has introduced some new questions. For example, can AI own a copyright? In March 2025, The Federal Circuit Court of Appeals held in Thaler v. Pearlmutter that only humans are eligible to be copyright authors (i.e., works created solely by AI cannot be copyrighted).
While that case addressed questions of authorship, it leaves open a new issue. What happens when AI infringesyour copyrighted work? The US Copyright Office states that infringement occurs when a work is reproduced, distributed, performed, publicly displayed, or made into a derivative work without permission from the owner. In other words, anytime the copyrighted work is copied and reproduced, that is copyright infringement – barring some exception. Most AI systems learn information that is given to it from articles, posts, and other sources. That information can then be used to generate content requested by users in the future. If a copyrighted work is directly copied, even if not entirely shared with users, that act of reproduction itself may constitute copyright infringement. In addition, if the work is shared with users, that is a separate act of distribution that could also give rise to copyright infringement.
The Chicago Artists Case
In December 2025, a group of Chicago artists and others sued Mureka, an AI platform that allows users to generate jingles, lyrics, and songs in seconds. Mureka works like any other AI platform and “learned” songs posted online and uses the music and lyrics of those “reference tracks” to generate new songs. So, here, there are multiple copyright rights violated creation of a derivate work (new songs), and also distribution of the new works.
Let’s start with what copyright infringement may occur from this type of use. One claim of infringement stems from Mureka asking its users to upload YouTube links for direct similarity and inspiration. Once a link is added, Mureka “stream-rips” the YouTube video, i.e., Mureka uses the song from the YouTube link to train its AI. That stream-rip is a copy that infringes already. In addition, Mureka keeps a downloaded file of the song in its database. This means that song can be used repeatedly to influence other outputs from the AI model, each one of which is a derivative work, which also violates copyright. Both acts happen without paying for access to the music.
While the court has yet to decide, the sections below outline how the court will determine if Mureka’s fair use defense has any weight. If the artists’ copyright claim is successful, this case could help pave the way for artists to further protect their works in the future. However, even if it is found to infringe artists’ copyrights, the court could hold that the AI company didn’t do anything wrong due to the doctrine of fair use and therefore not be held responsible for its actions.
Part of the problem is Mureka’s secrecy in exactly how the AI was trained and the extent of the copyrighted songs it was allowed to take (i.e., all or part of the songs). For example, Stan Burjek recorded vocals on 12 songs, including “This Road.” However, the case is still early in its filing and little discovery has occurred. It isn’t clear to the artists how much of these songs Mureka used, if the AI accessed them without permission, or if the YouTube link to the song was uploaded to the interface. However, the artists claim that no matter how Mureka accessed the music, or how much was copied, that any use is unlawful copyright infringement and would not fall under a fair use exception.
What Exactly Is Fair Use?
“Fair use” is a defense to copyright infringement that legally allows various uses of copyrighted works without the owner’s permission. Examples in the statute of acts that might qualify include teaching, news reporting, or critiques. This should make sense. These uses serve the public interest in education and free speech. Plus, who would ever give permission for negative things to be said about their work?
Determining whether something is “fair use” is an often unpredictable determination that requires analyzing multiple factors. The court analyzing the Mureka case will have to walk through these factors including, but not limited to, the purpose and character of how the copyrighted work is being used, and the resulting market effects. Would it be possible for the court to rule that Mureka’s AI model’s use of copyrighted songs was fair?
1. Purpose and Character of the AI Use
The first question the court will ask in the Mureka case is what is the purpose and character of the AI’s use. Purpose and character are determined by the commercial use and transformation of the work. If the use is considered “fair” then it will be allowed to continue to be trained with copyrighted works that are pulled into the model when generating new content.
i. Commercial Use
The Mureka court will have to ask whether taking other copyrighted works off the internet and using them to train AI constitutes commercial use. While Mureka’s platform used to be free to use, Mureka now requires paid subscriptions to access its features. Since Mureka is profiting from its use of the copyrighted works, Mureka’s use is commercial.
ii. Transformation of the Work
The court will next ask how the work has been transformed, if at all. Transformation can be satisfied in one of two ways: 1) the generated content is a transformative work, or 2) the generated work is a transformative use. If the artists show that Mureka did not transform the work or the use, this factor will weigh in favor of the artists and against a finding of fair use. “Transformative work” means that the new creation is so different that it would never replace or supplant the original. For example, a work is transformed if someone were to be inspired by a vase in a classical painting and created the vase itself out of clay. These works are very different such that one does not supplant the other.
In the Mureka case, the Chicago artists have argued that the AI generated works can be highly similar to their originals and therefore, Mureka has completely supplanted their works. Since the data and information about the artists’ songs are stored in Mureka’s databases, their music can be used to produce sounds that are substantially similar. And when specific songs, like “Lights on Our Faces” are uploaded to Mureka and stored in the database, a substantially similar song is easily reproduced.
Additionally, the second possibility of transformation asks whether the work’s use is transformative. Here, the court would ask whether the Mureka-generated songs add different meaning, value, or purpose to the original songs. In this case, the purpose of the Chicago artists and Mureka alike is for people to listen, enjoy, and even share the music with others. No additional meaning or value is added to the originals and the uses are identical. Therefore, since there is no transformative work or use, as well as commercial use, this factor weighs against a determination of fair use.
2. Impact on Markets
Often the heart of the analysis for the impact on the market is whether the content would be a substitute of the original. This might not be true all the time, as major fans of a specific artist’s work will continue to listen to the music they create. However, here, Mureka absolutely supplants the original artists’ works when it creates highly similar music. Additionally, the core problem with a platform like Mureka is that with every song generated, artists are losing opportunities to charge listeners or license their music. When their music is stream ripped from YouTube, artists are also losing out on ad revenue that would be generated if someone looking for the song listened to a few seconds of ads first. Artists typically wish to be compensated for allowing others to use their work. Thus, because there would be a large impact on the market, this factor weighs against fair use.
Weighing the Factors
Overall, these factors do not weigh in favor of Mureka and therefore should not constitute fair use. The Chicago artists have presented compelling reasons that give room for the court to hold in their favor as the case continues. Mureka as a platform circumvents paying artists, copies their legally protected works, and then gives it out to the public for free. These actions should not be tolerated, let alone rewarded with continued engagement.
Why Does this Matter?
The claim against Mureka can feel like a drop in the bucket compared to all the AI-related litigation filed in the past few years. In January 2025, the New York Times filed suit against ChatGPT and Open AI for copyright infringement. In that case, the New York Times claimed that Open AI infringed its copyright when Open AI’s model accessed articles to “learn” the information without paying for it. While the New York Times has sued for damages, monetary relief alone cannot fully restore the loss of protection. After all, the AI has already learned the copyrighted information and shared it with others. The New York Times and Mureka cases are still pending. These cases will shape how courts will rule on AI copyright infringement cases in the future and determine whether artists can protect their works against AI. The case against Mureka, in particular, has potential to be a blueprint for future caselaw.
Emily Miller
Assistant Blogger
Loyola University Chicago School of Law, J.D. 2027