How should the court rule in RIAA vs Suno and Udio?

The big record labels sued two AI companies, Suno and Udio, for using their music libraries to train their large language models (LLMs) for copyright infringement. This case could have major implications on the AI industry, the music industry, copyright owners, and copyright generally. So how should the court rule?

As background, the music industry has carved up rights in some interesting, and confusing ways. When you listen to a song, you are listening to many licenses that may be held by multiple people: the composer and lyricist for writing it, the artist for performing it, and the Record Label for recording it. These Record Label’s rights are in the master, including public performance and synch of that master. Essentially, other than concert performances, the Label owns the rights to recordings and the sale and use of those recordings.

There are two ways to get the rights to use copyrighted content like music: paying for a license or through “fair use.” We have seen a lot of activity on licenses, particularly by the big players, Google, Microsoft and OpenAI. But for the smaller players, paying license fees before knowing if there is a market for its AI offering is not feasible. So many of these companies rely on a fair use argument. A fair use determination is a four-factor test:

1. What is the purpose and character of the use – is the use transformative? Is it “commercial” in nature?

2. How creative is the original work?

3. What percentage of the work was used?

4. What is the effect on the market for the original work?

We can look at these four factors and we can debate who has the better arguments, particularly for the purpose and character of use and the effect on the market. However, it is important to look at the case through a much larger lens. The United States is at the forefront of innovation and disruptive technologies. Ruling that AI companies must secure licenses that most could not afford would stifle this nascent industry and harm our ability to lead the charge. 

Additionally, the copyright framework is a national framework for a reason. Having a patchwork of court decisions out of the various circuits will only cause far more chaos and confusion for far too long – it would take a Supreme Court ruling to unify the courts and that would take years. 

The music industry has a bad track record when it comes to innovation. They fight hard to protect the status quo. They did not learn their lessons from Napster, MP3 players and digital streaming. The lesson was, don’t fight - adapt. But here they continue to fight for the status quo. This case is not about the use of these existing libraries, although that is what RIAA would have you believe. This case is about creating music using AI, which music has little or no copyright protection. If there are no copyrights, then no licenses are needed. If no licenses are needed, then there is no market and the Labels’ revenue streams are decimated. 

So how should the court rule? In favor of the status quo that would crush innovation and hurt the US’s pole position on disruptive technologies? No. First the Court should delay any ruling as long as possible to allow the various Federal Agencies to offer the best solutions. But barring that, any ruling should be as narrow as possible to allow for innovation while protecting the artists and the content creators who would truly be hurt by unfettered use of their creations.