AI Made It, So Nobody Owns It: Copyright Uncertainty in AI-Generated Music

The rise of AI music generation tools has created a tempting narrative: if a machine made it, nobody owns it, and therefore anyone can use it freely. This framing is appealing in its simplicity, but it misrepresents the actual legal landscape. The truth is more nuanced, more uncertain, and more consequential for indie musicians than a catchy slogan suggests.

The Human Authorship Requirement

Copyright law in most jurisdictions has long required human authorship as a foundational element. In the United States, this principle traces back to the Copyright Act of 1976, which protects “original works of authorship.” Courts have consistently interpreted “authorship” to require human creative input.

This does not mean that any involvement of non-human processes automatically bars copyright. A photographer using a camera does not lose copyright simply because the camera captures light mechanically. The question is degree: how much human creativity went into the final work?

The challenge with AI-generated music is that this question often lacks a clear answer. When you type “upbeat indie folk song about summer” into an AI tool and receive a finished track, the level of human creative control is genuinely debatable.

U.S. Copyright Office Guidance

In February 2023, the U.S. Copyright Office issued guidance on the registration of works containing AI-generated material. This guidance, while not binding law, represents the agency’s current administrative position.

The key principles are straightforward:

• Works that are entirely generated by AI without human creative input are not registrable.
• Works that combine human authorship with AI-generated material may be registrable, but only for the human-authored portions.
• Applicants must disclose the use of AI in their registration materials.

The Copyright Office applied this reasoning in several notable cases. In one widely discussed example, the graphic novel Zarya of the Dawn received registration for its text and the arrangement of images, but not for the individual AI-generated images themselves. While this case involved visual art rather than music, the underlying principle applies across media.

For music specifically, the Copyright Office has not issued detailed guidance. This means indie artists are left to interpret general principles in an area where those principles have not been tested.

The Spectrum of AI Involvement

It is helpful to think of AI involvement in music creation as a spectrum rather than a binary:

More likely to receive copyright protection:

• A song you wrote entirely by hand, then used AI to generate a single background texture
• A composition where you wrote the melody, lyrics, and chord progression, then used AI to generate accompaniment
• A track where you recorded live performances and used AI only for mixing or mastering suggestions

In a gray area:

• A song where you provided detailed prompts and iteratively refined the AI’s output, making creative selections at each stage
• A composition where you wrote some elements and AI generated others in roughly equal measure
• A track where you significantly altered AI-generated material through editing, arrangement, and effects processing

Less likely to receive copyright protection:

• A complete song generated from a single brief prompt with no subsequent human modification
• A track where the AI tool made nearly all creative decisions regarding melody, harmony, rhythm, and arrangement
• Music generated automatically without any human selection or curation

The boundaries between these categories are not legally defined. Until courts or legislatures provide clearer rules, indie artists must make judgment calls in uncertain territory.

Why “Nobody Owns It” Is an Oversimplification

The claim that AI-generated music is automatically in the public domain overstates what the law actually says. Several complications exist:

The AI tool’s terms of service may matter. Many AI music platforms grant users a license to the outputs. Whether these licenses are enforceable is itself uncertain, but they create a contractual framework that exists alongside copyright law.

The training data raises separate questions. Even if the AI-generated output lacks copyright protection, the process of creating that output may have involved copying copyrighted works. This creates a different set of legal issues related to the training process.

Derivative works are complicated. If you take an AI-generated melody and build a full arrangement around it, the arrangement itself may be copyrightable even if the underlying melody is not.

Jurisdictional differences create additional uncertainty. The EU and UK are developing their own approaches to AI and copyright, which may differ from the U.S. position in important ways.

The practical reality is that “nobody owns it” is not a reliable legal conclusion. It is an assumption that has not been tested in court and may not survive legal challenge.

Registration Challenges

For indie artists who want to register their AI-assisted works with the U.S. Copyright Office, the process involves several practical hurdles:

Disclosure requirements. You must identify which portions of your work contain AI-generated material. This can be difficult when the boundaries are unclear.

Limiting your claim. You may need to disclaim the AI-generated portions, which limits the scope of your registration and potentially your ability to enforce the copyright.

Examination uncertainty. Copyright Office examiners may reach different conclusions about the same work, leading to inconsistent outcomes.

Cost of appeals. If your registration is refused, appealing the decision involves time and potentially legal fees that many indie artists cannot afford.

These challenges do not mean registration is impossible, but they do mean it requires careful preparation and realistic expectations.

Varying International Approaches

The legal treatment of AI-generated works differs significantly across jurisdictions:

United States: The Copyright Office currently requires human authorship and has rejected registration of purely AI-generated works. The approach focuses on the degree of human creative control.

United Kingdom: The UK’s Copyright, Designs and Patents Act 1988 includes a provision for “computer-generated” works, which may offer protection for AI-generated music. However, this provision has not been extensively tested in the AI context, and its applicability remains debated.

European Union: The EU generally requires human authorship, but individual member states may interpret this differently. The EU AI Act focuses primarily on transparency and disclosure rather than copyright ownership.

Other jurisdictions: Countries like Japan, China, and Australia are developing their own approaches, creating a patchwork of rules that may affect where you choose to release or register your work.

For indie artists working internationally, these differences create practical complications. A work that receives copyright protection in one country may not receive it in another.

Common Misconception

“AI-generated music is automatically public domain” — The legal status is genuinely uncertain. While purely AI-generated works may lack copyright protection in some jurisdictions, the boundaries are unclear, and many AI-assisted works involve enough human creativity to potentially qualify for protection.

What This Means for Indie Artists

If you are using AI tools in your music creation process, several practical considerations apply:

Document your creative process. Keep records of your prompts, iterations, edits, and creative decisions. This documentation may be valuable if you need to demonstrate human authorship.

Consider the spectrum. The more human creativity you contribute, the stronger your potential copyright claim. If ownership matters to you, prioritize human creative input.

Read the terms of service. Understand what rights you have to the AI tool’s outputs under your user agreement.

Be cautious about relying on copyright. If you release music that is substantially AI-generated, do not assume you can enforce copyright against others who use it.

Stay informed. This area of law is evolving rapidly. Guidance that applies today may change as courts and legislatures address these issues.

Practical Checklist

• Document your creative process from prompt to final output
• Review your AI tool’s terms of service regarding output ownership
• Consider how much human creativity you contributed to the final work
• Be cautious about registering works that are substantially AI-generated
• Monitor legal developments in your jurisdiction
• Consult a lawyer if copyright ownership is critical to your project

When to Ask a Lawyer

• You are releasing a commercially significant work with substantial AI involvement
• You need to register a copyright and are unsure how to disclose AI use
• Someone has copied your AI-assisted music and you want to enforce your rights
• You are entering a licensing deal where ownership questions matter

Sources

1. U.S. Copyright Office, “Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence” (February 2023)
2. U.S. Copyright Office, “Copyright and Artificial Intelligence, Part 1” (Notice of Inquiry, August 2023)
3. U.S. Copyright Office, “Zarya of the Dawn” Registration Decision (February 2023)
4. UK Intellectual Property Office, “Consultation on Artificial Intelligence and Copyright” (2023)
5. European Parliament, “EU Artificial Intelligence Act” (2024)