If you’re creating work with AI tools—whether you’re generating images, writing code, composing music, or blending human effort with machine output—you face a unique challenge: proving that you are the creator. The U.S. Copyright Office now requires creators to disclose AI-generated content in registration applications and to document exactly what you contributed. Without clear evidence of your creative control, you risk losing copyright protection entirely. This guide shows you how to document your authorship so you can defend your work if a dispute arises.

Why AI Authorship Documentation Matters Now

Copyright protection has always rested on a simple principle: a work must be the product of human authorship. But when AI tools generate content in response to a prompt, the line between human creativity and machine execution blurs. The U.S. Copyright Office has taken a clear stance on this issue. According to federal copyright registration guidance issued in March 2023, when an AI technology receives only a prompt from a human and produces complex written, visual, or musical works in response, the “traditional elements of authorship” are determined and executed by the technology—not the human user.

This means that purely AI-generated content cannot be copyrighted. But works containing both human authorship and AI-generated material can still receive copyright protection—if you document your contributions properly. The Office recognizes that a human may select or arrange AI-generated material in a sufficiently creative way that “the resulting work as a whole constitutes an original work of authorship,” or modify AI material to such a degree that the modifications meet the standard for copyright protection.

The key word is document. Without evidence of what you actually did, the Copyright Office will assume the AI did it all, and your registration will be refused or canceled.

Understanding the Copyright Office’s AI Disclosure Requirement

The U.S. Copyright Office now imposes a duty on applicants to disclose the inclusion of AI-generated content in any work submitted for registration. This is not optional, and failure to disclose—or misrepresenting AI content as human-authored—can result in cancellation of your registration or loss of copyright protection in court.

What You Must Disclose

According to the Copyright Office guidance, AI-generated content that is more than de minimis (more than trivial) must be explicitly excluded from your copyright claim. You do this by describing the AI-generated portions in the “Material Excluded” section of your registration application, using language like “[description of content] generated by artificial intelligence.”

At the same time, you must clearly identify what you created. In the “Author Created” field, you should describe the authorship you contributed. For example:

  • If you incorporated AI-generated text into a larger work, claim the portions that are human-authored.
  • If you creatively arranged human and AI content, write: “Selection, coordination, and arrangement of [describe human-authored content] created by the author and [describe AI content] generated by artificial intelligence.”
  • If you modified AI output substantially, describe your modifications and the reasoning behind them.

Do not list the AI technology or the company that provided it as an author or co-author. The tool is not a creator—you are.

What Happens If You Don’t Disclose

If the Copyright Office discovers that you failed to disclose AI-generated material, it may take steps to cancel your registration. Separately, if you knowingly provided inaccurate information to the Office, a court may disregard your registration in an infringement action if the accurate information would have resulted in refusal of the registration. In other words, non-disclosure doesn’t just cost you the registration—it can cost you the ability to enforce your copyright in federal court.

Documenting Your Creative Control: A Practical Framework

The Copyright Office’s test is whether you contributed more than a trivial variation to the final work. To pass that test, you need contemporaneous evidence—records made during the creation process, not reconstructed after the fact. Here’s how to build that evidence:

Preserve Your Human-Authored “Seed” Work

Before you use any AI tool, save and timestamp your initial human input. This might be a sketch, a draft, a photograph, a code skeleton, or a written outline. This seed work serves two purposes: it proves you had an independent creative starting point, and it gives you a fallback claim if the downstream AI output is later disclaimed or refused.

Save your seed as a fixed artifact with a clear date and your signature or identification. Use version control systems (like Git with signed commits), scanned and dated PDFs, or numbered drafts. Keep this seed isolated from later AI-modified versions so the distinction is unmistakable.

Document Each Human Modification with Reasoning

For every change you make to raw AI output, record what was changed, why you changed it, and what alternative you rejected. In the SURYAST case, the Copyright Office Board attributed expressive choices to the AI because nothing in the record showed any modifications by the applicant. A line-by-line record of edits and the reasoning behind them is the most direct evidence that you contributed more than trivial variation.

Save before-and-after versions of every AI output that you edit. Annotate the differences with the reason for each change and any alternative you considered. Capture this record in the same commit or document version as the change itself, not in a separate retrospective summary written months later.

Record Selection, Arrangement, and Structural Decisions

Even where individual elements are AI-generated, the Copyright Office recognizes that human selection, coordination, and arrangement of that material can supply the traditional elements of authorship. Document the human choices about organization, ordering, hierarchy, and integration.

For example:

  • Document file and module organization, naming conventions, and integration choices.
  • If you selected AI-generated candidates from a larger pool, log which candidates you considered and why you chose each one.
  • If you rearranged or reordered AI-generated elements, record the original order and your rationale for the change.
  • If you combined multiple AI outputs into a coherent structure, document how and why.

Building Tamper-Evident Records: Copyright Provenance

Documentation is only useful if it can be verified and trusted. Copyright provenance captures the iterative timeline of creation—the decisions, contributions, and tools used along the way—and makes that timeline tamper-evident and independently verifiable. This is distinct from a one-time copyright registration; it’s an ongoing record of how your work came to be.

For creators working with AI tools, copyright provenance should include:

  • DAW session data (for music creators): Save and timestamp your digital audio workstation sessions showing which tracks, effects, and decisions were yours.
  • Version histories: Use Git, GitHub, or similar version control to create an immutable record of every change, with commit messages explaining your reasoning.
  • Attribution records: Document which parts of the work came from AI and which came from you.
  • Prompt logs and workflow evidence: Keep records of the prompts you used, the AI outputs you received, and the modifications you made.
  • Timestamps and signatures: Use RFC 3161 timestamping or notarization to create independent verification that records were made at a specific time.

This kind of evidence becomes critical if your copyright is ever challenged. In an AI-authorship dispute, a defendant may challenge validity through public statements, metadata, source files, version histories, prompt logs, or other workflow evidence obtained in discovery. A company’s future position may depend less on broad claims that AI was merely a “tool” and more on whether it can show, with contemporaneous evidence, what a human actually contributed to the authorship of the work.

Correcting the Record After Registration

If you’ve already obtained a copyright registration and later realize you failed to disclose AI-generated content, you can correct the public record by submitting a supplementary registration. This is a special type of registration used to correct an error in a copyright registration or to amplify the information given in a registration.

In the supplementary registration, you should:

  • Describe the original material that the human author contributed in the “Author Created” field.
  • Disclaim the AI-generated material in the “Material Excluded/Other” field.
  • Complete the “New Material Added/Other” field with any additional information.

As long as there is sufficient human authorship, the Copyright Office will issue a new supplementary registration certificate with a disclaimer addressing the AI-generated material. This corrective step is far better than leaving an inaccurate registration on the public record.

Conclusion: Document Now, Defend Later

Copyright protection for AI-assisted works is not automatic, and it’s not guaranteed just because you used an AI tool as part of your creative process. What matters is what you did, and whether you can prove it. The U.S. Copyright Office’s disclosure requirement and the evolving case law around AI authorship make it clear: creators who work with AI must document their contributions contemporaneously, preserve their seed work, record their modifications, and maintain tamper-evident records of their creative decisions.

If you’re using AI tools in your creative work, start building this evidence today. Save your drafts, document your reasoning, timestamp your versions, and keep clear records of what the AI generated versus what you created. When you’re ready to register your work with the U.S. Copyright Office, you’ll have the evidence you need to make a credible claim of human authorship. And if your copyright is ever challenged, that documentation will be the foundation of your defense.