Businesses are using generative AI to draft marketing copy, produce illustrations and speed up creative work. Yet one question follows those projects: can the finished material be protected by copyright? Under current US law, the answer turns on what a human contributed. The latest ai-generated content copyright rules do not create a blanket ban on AI-assisted work. They distinguish human expression, which may qualify for protection, from material whose expressive elements were produced by a machine.
The Current US Rule: Copyright Still Requires a Human Author
The US Copyright Office says copyright protects original human expression. In its January 2025 report on copyrightability, the Office concluded that existing copyright principles can address generative AI without a new category of protection for machine-produced outputs. AI may be part of the creative process, but it cannot be the legal author of a work by itself.
That position has judicial support. In 2025, the US Court of Appeals for the District of Columbia Circuit upheld the refusal to register an image that listed an AI system as its sole author. The US Supreme Court declined to review that decision on March 2, 2026. A work claimed as entirely machine-authored therefore cannot currently receive US copyright registration simply because a person owns or operates the AI system.
This does not mean every image, article or design involving AI is unprotected. The key question is whether identifiable parts of the finished work reflect enough original human creativity.
How Much Human Involvement Is Enough?
There is no universal percentage test. The Copyright Office evaluates each work individually and looks at who determined the copyrightable expressive elements. A person may qualify as an author by creating protectable material, creatively arranging content or substantially modifying AI output.
Prompts Alone Usually Do Not Establish Authorship
Under current US Copyright Office AI guidance, entering prompts is generally not enough by itself. Even a detailed prompt may describe a desired result without giving the user reliable control over the precise expression produced. The model may decide the exact wording, composition, lighting, colours or visual relationships.
This distinction matters for AI content copyright. Effort, time and technical skill do not automatically create copyright. The law focuses on original human expression, not merely the labour involved in requesting an output. As AI systems develop, the analysis may evolve, but businesses should not assume that a long prompt automatically makes the result protectable.
Human Editing, Selection and Arrangement Can Matter
Protection is more likely when a person goes beyond accepting the first result. A designer might redraw major portions of an AI image, combine elements into a new composition and add original typography. An editor might use AI suggestions but write substantial passages, restructure the narrative and make original choices about tone.
Copyright may then cover the human-authored additions, modifications, selection or arrangement. It does not necessarily extend to the underlying AI-generated elements. A mixed work can contain both protected and unprotected material.
Who Owns AI Art and Other Generated Material?
AI art ownership is often discussed as though the person who typed the prompt automatically owns a copyright. That is too simple. If the output lacks sufficient human authorship, there may be no copyright in it to own under federal law. A platform’s terms may give a user contractual rights to use generated material, but a contract cannot turn uncopyrightable machine expression into a federally protected work.
Where a human makes copyrightable contributions, ownership generally begins with that author, subject to rules involving employment, work-made-for-hire agreements and written assignments. Businesses should document who created the human elements and ensure employee or contractor agreements address intellectual property rights.
Registering a Work That Contains AI-Generated Material
Applicants must be accurate about AI involvement. The Copyright Office requires disclosure of more than minimal AI-generated content and expects applicants to identify the human-authored material being claimed. Unclaimable AI-generated material should generally be excluded, while the application can describe protectable contributions such as original text, artwork, editing, selection or arrangement.
A company should preserve drafts, source files, revision histories and notes showing how a person shaped the final work. These records do not guarantee registration, but they can help explain the creative process. If an earlier application omitted significant AI-generated material, the registrant may need to correct the record.
Copyrightability Is Only One Part of Generative AI Legal Risk
A work can be unprotected by copyright and still create legal problems. Generative AI legal risk may involve alleged infringement, trademark confusion, publicity rights, privacy, defamation, confidential information or contractual restrictions. Businesses should review outputs before publication rather than treating an AI tool as a legal clearance system.
Training Questions Are Different From Output Ownership
The debate over whether AI developers may train models on copyrighted works is separate from whether a particular output has human authorship. The Copyright Office has examined training, licensing and potential liability separately. Businesses should not confuse those disputes with the copyrightability of their own contributions.
Similarity and Brand Risk Still Require Review
Teams should check whether an output closely resembles protected characters, photographs, artwork, logos or distinctive branding. They should also avoid placing confidential information or personal data into tools without understanding the provider’s terms and data practices. Human review remains essential for commercially important content.
A Practical AI Copyright Policy for Businesses
An internal policy should record tools used, require human review, preserve evidence of human creation and set approval rules for high-value content. It should clarify when legal review is needed, especially for advertising, product designs, entertainment projects and assets intended for licensing.
For valuable works, build human authorship into the workflow from the start. Use AI for brainstorming or assistance, then have creators make original expressive decisions, substantially revise the material and document those changes. This cannot eliminate uncertainty, but it gives the business a stronger basis for claiming copyright in the human-created portions.
Frequently Asked Questions
Can I copyright an image created entirely by AI?
Generally, no. If the expressive elements were generated entirely by the system without sufficient human authorship, the image is not eligible for US copyright protection under the current standard.
Does editing AI-generated content make it copyrightable?
It can. Original human edits, additions or creative arrangements may be protected when they are sufficiently expressive. Copyright would normally cover those human contributions, not every machine-generated element.
Do I have to disclose AI use in a copyright application?
You should disclose appreciable AI-generated material and identify the human authorship being claimed. Minor or de minimis AI involvement may not require exclusion, but accuracy is essential.
Can my business use AI output if it has no copyright?
Possibly, but lack of copyright protection does not make use risk-free. Platform contracts, third-party copyrights, trademarks, publicity rights, privacy obligations and other laws may still apply.
Conclusion
The ai-generated content copyright rules in the United States are built around a straightforward principle: copyright protects human creativity, not autonomous machine authorship. AI can assist a creator without destroying protection, and a mixed work may qualify when a person contributes original expression through writing, editing, arrangement or modification. Businesses should treat AI as a tool, keep humans responsible for meaningful creative decisions and maintain records showing what those humans contributed.
