Music That Is Entirely AI-Generated Cannot Be Copyrighted, but Who Owns an AI-Assisted Song?

Copyright laws must be obeyed and enforced to prevent damaging situations for the future of the music industry.
Music That Is Entirely AI-Generated Cannot Be Copyrighted, but Who Owns an AI-Assisted Song?
Like

Share this post

Choose a social network to share with.

This is a representation of how your post may appear on social media. The actual post will vary between social networks

The use of AI in music has sparked considerable debate, particularly regarding lyric composition and the generation of new melodies. While entirely AI-generated music cannot be copyrighted, the circumstances are different for AI-assisted songs.

Understandably, human creativity deserves protection. There is a significant difference “between pure, human art and bot-influenced art,” Bloomberg shared in an article. AI not only lacks originality, but it also draws ideas from pre-existing content.

The Recording Academy has raised concerns about using copyrighted material to train AI models. “Owners must have a right to decide when and how their work is used,” writes Montana Miller, from the Academy’s Advocacy and Public Policy Office. The rapid development of AI technology makes this difficult, as unsolicited use of material is often done without the consent or knowledge of the copyright owner, despite legal protection being in place. Copyright laws must be obeyed and enforced to prevent damaging situations of copyright infringement.

Recently, The Verge reported that Universal Music Group established a process through the Digital Millennium Copyright Act (DMCA) for “detection tools to submit a takedown request.” This development occurred after the AI-generated track “Heart on My Sleeve” featuring Drake and The Weeknd went viral. There is speculation that UMG did not take the internal steps to contact streaming services to enforce the track’s takedown. The fact that these straightforward scenarios have allowed musicians to be put at risk is concerning for the future of the recording industry.

The Basics of Music Royalties 

Let’s break down music royalties into two of their main forms:

Master / Sound Recording

The master copyright determines ownership of a sound recording. A musician who owns their master rights in full would receive 100% of the master's earnings. However, this share varies if other individuals or entities, such as a producer or record label, were involved in the creation process. AI-assisted songs introduce an unclear distinction of ownership since audio is more difficult to trace. AI might provide a sample for someone to work from, which they can recreate independently. How do we track what percentage of a melody came from AI-generated ideas? Should unclaimed royalties go to waste?

Composition / Songwriting

The level of AI assistance in a song's composition or lyrics can vary, from suggesting topic ideas and titles to writing an entire verse or song. The U.S. Copyright Office urges applicants to disclose any part of a work that involves the use of AI. However, copyright protection continues to be granted for many AI-assisted works, which include “about 200 registrations of the 1,000 applications the office has received that disclose the use of AI.” There are also major concerns around instances of AI detection software falsely claiming ownership.  Currently, the primary method for identifying AI-assisted songs relies on artist transparency. There must be a more viable solution for determining the ownership in these cases.

Who should own the rights of AI-assisted songs? Is there a solution that can accurately track a musician’s authenticity and ownership?

Please sign in

If you are a registered user on AVIXA Xchange, please sign in

I'm curious to see what comes of this, especially as major audio AI generation platforms are getting into legal battles with major labels as a result of a lack of clarity. I know online platforms are licensing their content to genAI companies in order to secure more revenue, but I don't think that'll work here.