As of late, generative AI systems have become the hottest musical artists, as they’ve raised legal questions related to copyright law.
In April, such a system authored a song titled “Heart on My Sleeve,” which featured voices exactly like those of Drake and The Weeknd. On multiple streaming platforms, the track received millions of listens, while confusing fans with its verisimilitude to recognizable voices. In response to the widespread digital consumption of the track, Spotify, Apple Music, YouTube, and TikTok pulled the song from circulation.
Ghostwritten Songs
Attributed to a social media and music self-publishing platform user, known only as “Ghostwriter,” “Heart on My Sleeve”has raised many concerns about intellectual property. It has even shaken historical understandings of such property as a legal concept, as generative AI is being taken more and more seriously amongst some of the most powerful organizations in music.
One such organization is Universal Music Group, the international music company that represents artists like Drake and The Weeknd. They were not having any of Ghostwriter’s antics. In the wake of the release of “Heart on My Sleeve,” attorneys for Universal sent a letter to streaming services requesting that such platforms block AI systems from accessing their material to learn the elements of copyright-protected music.
Other users on various streaming and social media platforms have also used artificial intelligence, much like Ghostwriter. For example, a TikTok user, known only as “garytheproducer,” used generative AI to recreate Ariana Grande’s voice in a cover of Drake’s “Controlla.” The cover has received well over one million views and over 100,000 likes.
Given the response from Universal to the musical artistry of generative AI, it’s pretty clear that such AI systems are raising red flags amongst even the biggest players in the music industry. Wherever you turn, it’s becoming clearer that AI is a force to be reckoned with. A revolution has begun, and it’s changing the way we understand everything—including what it means to be a creator in the realm of music.
Some artists have been more friendly to the AI, however. After all, cooperating can be lucrative; in theory, it’s not that different from another artist sampling your song–for which you get royalties. In an interview with Rolling Stone, the artist known as Grimes announced that anyone could use her voice in an AI-generated song, as long they pay her 50% of the royalties made from such material.
AI Raises Philosophical Questions
Under U.S. copyright law, the reproducer of any original work will almost always need a copyright license to distribute reproductions of those works.
The Harvard Business Review has weighed in on these matters in their April article, “Generative AI Has an Intellectual Property Problem.” The authors draw attention to a 2022 case called Andersen v. Stability AI. In that case, artists filed a class action lawsuit against AI platforms. In the complaint, the artists claim that the platforms’ uses and reproductions of their original material constitute copyright infringement.
In the Anderson case, the crucial issue really is one of who actually owns the rights to the material alleged to have been infringed by the AI systems. Similar questions are also raised by the songs released by Ghostwriter and garytheproducer; music from the two bears far too close a resemblance to that of other famous artists. But do works generated by AI systems really constitute copyright infringement, in the ways alleged in the Anderson case and those of other generative AI music creators?
Generative AI learns from material to which it is exposed. It draws from such material. As a consequence of its learning processes, it creates and engages in intelligent activities in ways that at least appear to be the same as those of human beings. In the human process of engaging in an act of creation, even a work understood to be “original” by law, human beings also engage in what appears to be a process that is remarkably similar to those of generative AI systems engaged in such processes. As such, isn’t an AI system like a human creator, in that it draws upon experience to engage in productive activities—whether it be within the scope of art or in some other area of life?
Beyond the philosophical question of what it means to be human and to engage in the act of creating, there are of course tougher legal questions raised by AI.
Does AI Violate Copyright Law?
Under circumstances like those of Ghostwriter and garytheproducer, monumental questions concerning intellectual property laws are stake.
Article 1 of the U.S. Constitution reads that Congress shall have the power “[t]o promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries.” This gives creators sole rights to reproduce their work for defined periods of time.
Meanwhile, the U.S. Copyright Office indicates that copyrights protect “original works of authorship including literary, dramatic, musical, and artistic works, such as poetry, novels, movies, songs, computer software, and architecture.”
In situations like those involving Ghostwriter and garytheproducer, it’s pretty natural for someone to ask: who actually owns the rights to the music, and what does it mean for something to be “original?”
The authors of the Review article say that the court must consider whether the platforms are liable for “unauthorized” and “derivative” works. If the material on the platforms is determined to be “insufficiently transformative” from the artists’ copyright-protected works, the court will likely decide that the platforms have committed infringement. In other words, at issue is whether AI-generated reproductions of these artists’ material are different enough from these artists’ material that they do not constitute infringement.
The authors also argue that how the courts will decide Andersen hinges on how the fair use doctrine is interpreted in the case. Under some circumstances, this doctrine allows copyrighted material to be reproduced without a license to do so. This doctrine permits reproductions without licenses when it comes to satire, commentary, reporting, teaching, and scholarly works. In the Anderson case, the court must also consider whether the products of the AI platforms constitute what the fair use doctrine permits as these types of reproductions where licenses are not required. If the court finds that the works on the platforms are such types of reproductions, the fair use doctrine would shield the AI platforms from liability for copyright infringement.
Take Aways
Remember, this case hasn’t been decided yet. It’s just another instance of uncertainty about how copyright law may change in a world where generative AI is capable of passing itself off as music superstars. Regardless of how the case turns out, it’s clear that protections that have historically been given to original works of authorship are being shaken to their very core.
Related Resources
For more information about laws and legal issues related to music created by generative AI systems, review FindLaw’s Learn About the Law pages:
- Is Your TikTok Video Violating Copyright Laws?
- What is the Fair Use Doctrine?
- How Long Does Copyright Protection Last?
The post Generative AI: Biggest Threat to the Music Industry Since Napster? appeared first on .
