Is AI the Future of Music? Legal Implications and Industry Predictions

This summer, Belgian singer Angèle responded to her own voice being used in a cover that has now exceeded 6 million views on YouTube. The cover, titled "Saiyan," originally featured the two French rappers, Gazo and Heuss l’Enfoiré. Comments on the cover ranged from "She should sing songs like this one; it's so good!" to "So much better than the original!" Angèle's reaction to this experiment was mixed, as her caption reflected: "I don't know what to think of artificial intelligence. I find it mind-blowing, but at the same time, I'm worried about my profession." In response to numerous requests from her fans, she finally performed the song at her concert four weeks ago, and the crowd went wild. 

 

This situation raises questions about the meaning of creating music autonomously and creatively. On one hand, AI presents a revolutionary tool that allows for unprecedented creativity and musical exploration. On the other hand, it also brings valid concerns and potential threats related to intellectual property and personality rights.

 

While the recent use of AI in music creation has sparked a significant shift, the roots of AI-generated music date back to 1957. During this time, American composer, Lejaren Hiller, and American mathematician, Leonard Isaacson, programmed a computer to generate various musical elements, including pitches, rhythms, and dynamics. The resulting composition, "Illiac Suite for String Quartet", while not a dancefloor hit, laid the foundation for what would follow: the ability to emulate popular artists. 

 

 

* * *

 

Some artists are concerned about the economic loss but also the ethical implications of their voices being used without consent, as exemplified by Robin Williams' daughter's comments on how disturbing it was hearing her late father's voice. 

 

What Can Artists Do to Prevent AI Music Using Their Voice/Songs?

 

1.     Flagging Copyright Infringement : What Can and Can't Be Copyrighted in a Song? 

The track "Hard On My Sleeve" sparked conversations as it remarkably resembled a collaboration between Drake and The Weeknd. Created by Duo Alta, consisting of American rapper Mr. J. Madeirus and French producer Van Silisup, it demonstrated AI's ability to mimic renowned artists' distinctive voices.

While the song was gaining popularity, Universal Music Group successfully had it removed from streaming platforms. It is worth noting however that the takedown notice issued by Universal targeted the sample used in the song, not the voices themselves. Indeed, voices are generally not considered copyrightable since they do not meet the definition of "original works of authorship fixed in a tangible medium of expression."  As  the voices of Drake and The Weeknd were not covered by copyright, Universal's strategy for recognizing infringement focused on the unauthorized use of a protected sample. 

This appears to be a solution for affected artists: takedown notices should be directed at parts of the song that genuinely infringe upon copyrighted works, such as the musical composition or the sound recording.

 

2.     Claiming a violation of personality rights.

 

Personality rights are typically characterized by two primary rights: the right of publicity and the right to keep one's image and likeness from being exploited for commercial purposes without consent or proper compensation. 

 

Although voices can't be copyrighted, artists retain exclusive commercial rights to their names, ensuring that songs cannot be attributed to them without consent.

 

However, establishing a commercial purpose may require proof, and in the case of AI-generated music, the commercial aspect can be ambiguous. Many individuals create AI-generated music and share them without direct commercialization. Does accruing views on platforms like YouTube and TikTok constitute a commercial benefit?  The question remains open. 

Is the AI Tool Infringing Artists' Rights?

 

The issue of whether AI tools are infringing IP rights centers on the balance between input and output in AI-generated music. Two key factors come to light: the data sets used to train AI systems and whether the output of generative AI infringes on the copyrights of the materials used in its training. 

 

A class-action suit was brought by various authors against OpenAI for using their copyrighted materials for AI training purposes. Daily newspaper The New York Times also considers legal action against the tech giant to protect the work of its journalists.  These actions bring up similar concerns in the realm of music. Indeed, to generate songs "in the style of" Drake, The Weeknd, Angèle, or Jay-Z, AI systems need to be trained using copyrighted materials from these artists, typically without their explicit consent. It remains to be seen whether musicians will follow the same path as writers and take action against the system rather than the end-users. 

Can AI Users Copyright The Song They Generate with AI? 

 

Central to these discussions is the issue of human involvement in AI-generated music. The US Copyright Office has recently introduced guidelines that call for a certain degree of human participation in the creative process (Refer to the article "Copyrighted by Robots? Exploring the Intersection Between AI and Creative Ownership" 

at https://thejurisfashion.blogspot.com/2023/04/copyrighted-by-robots-exploring.html).

This raises the question of whether AI should serve as an "assistant," meaning it plays a role in assisting the creative process, alongside a human contributor, for example, when an author provides part of a song and uses AI to finalize it.

While the Copyright Office's guidelines aim to address these challenges, they still carry a degree of ambiguity due to the complexities of each case. For example, does auto-tuning (the altering of the level of the voice pitch by means of a computer) satisfy the necessary human input requirement? 

 

Is AI the Future of Music? Industry Professionals Predictions

 

The future of music with AI is a topic of debate among industry professionals. Some predict that record companies and publishers will push artists and songwriters to transfer their rights, ultimately leading to entirely computer-generated music. This perspective highlights concerns about a potential loss of human involvement and compensation for creators.

Conversely, some predictions present a more balanced outlook. They acknowledge that record labels and music publishers may find ways to own and monetize AI-generated music while emphasizing the continued usefulness of AI as a creative tool. They also anticipate some form of copyright protection for generative AI, recognizing the complexities in disentangling AI-generated content. This perspective implies a coexistent relationship between AI and human involvement in the music industry.

 

As new laws and regulations surrounding AI-generated music are expected to emerge, legislators will face the delicate task of striking a balance between safeguarding the rights of existing artists and promoting artistic creation, the latter resulting in the introduction of countless new artists and a wave of fresh music.

ASR