Disclaimer: This article is published for informational purposes and does not constitute legal advice.
If we can learn anything from the history of the use of synthesizers, DAWs, and Autotune in music, it’s that it’s not worth trying to fight technological advancements in this industry. So if you’re wondering whether the use of generative Artificial Intelligence (AI) in the songwriting and production process will go away, chances are it’s here to stay.
Whether or how much AI should be incorporated in music is more than just a creative decision; it’s also a business decision. After all, the use of AI can have a deep impact on intellectual property.
In the United States, the discussion on copyrights related to generative AI is moving fast. As of October 2023, the U.S. Copyright Office (USCO) is studying copyright implications regarding generative AI to advise Congress on potential legislative changes. There are already a few guidelines for those who want to apply for copyright registration of works made with generative AI.
Copyright registration is not mandatory for a work to be worthy of protection, but these guidelines set out a solid framework for anyone dealing with music created with generative AI.
So how can musicians, songwriters, labels and/or publishers own publishing rights over music made with generative AI? The answer to this question depends on two main factors:
- whether there is any human involvement at all in the songwriting or production process.
- whether there is any legal requirement to copyright works made with generative AI.
Let’s look at things not to do, and things you should do when copyrighting AI music.
Don’t Credit Yourself If All You’ve Done Was Create The Prompt
It’s one thing to use AI to “create” or improve a few elements of the music-making process, and a very different one is to leave the entire process to the AI. In the second scenario, the outcome of the AI work is something that cannot be attributed to any person.
This is because copyright laws and court decisions require that the author of an original work should be human. A software won’t qualify as an author.
One may argue that creating a prompt should be enough creative input for an individual to be eligible for copyright over an AI-made song, but that’s not how the USCO understands.
Released in March 2023, the USCO’s Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence states that “users do not exercise ultimate creative control over how such systems interpret prompts and generate material.”
Don’t Credit The AI Developers Or Their Company
As an alternative to avoid copyright infringement lawsuits, some people may want to credit the company that owns the rights over the AI, the AI developers, or even the AI itself. Unfortunately, that’s no better than crediting the songwriter or producer operating the AI.
The USCO determines that “[a]pplicants should not list an AI technology or the company that provided it as an author or co-author simply because they used it when creating their work”.
Do Disclose The Use Of AI When Seeking Registration
Despite these don’ts, it’s not that music made with AI cannot be protected.
If there is human input together with AI input, a work may still be eligible for copyright protection.
If you’re going to submit a work for copyright registration towards the USCO, you should disclose that your work contains AI-generated content. The USCO’s Copyright Registration Guidance also states that applicants should “provide a brief explanation of the human author’s contributions to the work.”
Such disclosure can be as simple as stating: “[description of content] generated by artificial intelligence” If you’re unsure of what to disclose, you can “simply provide a general statement that a work contains AI-generated material” and the USCO will contact you to determine how to proceed.
The applicant that does not disclose the use of AI might be asked to provide details of their creative process. If they do not prove the existence of substantial human involvement, it may lead to refusal of registration, such as happened to artist Jason Allen when trying to copyright the artwork “Théâtre D’opéra Spatial”.
As of now, there are no landmark decisions by the USCO or U.S. Courts on AI music specifically. However, cases like Allen’s or Kris Kashtanova’s novel “Zarya of the Dawn” (whose copyright registration is at risk of being canceled) set important precedents for all creative industries.
Do Check On Other IPs And Rights
Composing and producing is one part of the job, but some AIs go as far as generating an entire track with vocals and even emulating the voices of famous singers.
The protection of a voice is not to be confused with copyrights. Still, the misappropriation of a voice and/or particular singing style can be a violation of personality rights, and you may not be able to use the track anymore.
That’s what happened when producer Ghostwriter created a Drake-soundalike track with AI. Universal Music filed a claim on YouTube and the track video was taken down. Even though that was a copyright claim, it was the right to use Drake’s voice (or rather the AI’s resemblance with it) that the label was going after. This is a clear case of a violation of rights other than copyrights that end up hurting potential income with publishing.
But there can be other copyrights involved in songwriting or producing with AI too — for example, when a producer commands the AI to sample a track. Unless the sampled track comes from a licensed library, there’s no difference here from when a sample is featured manually; clearance is needed likewise.
Whether U.S. laws will make it easier or harder to copyright AI music is something we’ll have to wait and see. What’s for sure is that we may expect new guidelines and requirements in the near future, since the topic is currently the subject of many studies from U.S. copyright authorities and scholars. We’ll be sure to keep updating on this hot topic and its impact in the music industry as and when developments occur.