Under current U.S. copyright law, art generated purely by artificial intelligence cannot be copyrighted, because the Copyright Office requires a human author. That single rule sits at the center of a growing legal and ethical debate: when an AI system produces an image, who, if anyone, qualifies as its author — the developer who built the system, the AI itself, or the person who typed the prompt? The answer determines who controls the work, who profits from it, and how human artists compete in a market increasingly flooded with machine-made images.
The Human Authorship Requirement in U.S. Copyright Law
The governing position comes from the U.S. Copyright Office itself. In registration guidance dated March 10, 2023, signed by Register of Copyrights Shira Perlmutter, the Office stated plainly that if a work's "traditional elements of authorship were produced by a machine, the work lacks human authorship and the Office will not register it." The Office gave a concrete illustration: when an AI technology receives solely 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." In the Office's assessment of currently available generative AI systems, users do not exercise ultimate creative control over how such systems interpret prompts and generate material.
The Office also left the door open, noting it "continues to monitor new factual and legal developments involving AI and copyright and may issue additional guidance in the future." In other words, this is the current interpretation of U.S. law, not a permanent constitutional fact — a distinction that matters as courts and lawmakers confront the technology.
Who Is the Author? Exploring the Candidates for AI Art Copyright
The legal debate, framed most thoroughly in a Houston Law Review article by Mackenzie Caldwell, asks which entity in the AI art pipeline fulfills the traditional requirements of an "author." Three candidates present themselves: the developer who created the system, the AI model that generated the image, and the end user who prompted it. Each candidate has a plausible claim, and each runs into a serious objection.
What makes the question genuinely difficult is that authorship in copyright law has never been a single fixed concept. Caldwell's analysis traces competing philosophical theories — including Kantian views of authorship as a "communication with the public" rooted in the autonomy of the author, contrasted with Lockean property theories focused on the physical labor of creation. Different theories point toward different answers about who deserves recognition, and current U.S. law has effectively resolved the tie by defaulting to a human authorship requirement.
The Developer as Author: Arguments and Limitations
The developer's claim rests on a simple analogy: they built the instrument that made the artwork possible, much as a camera manufacturer enables a photograph. But that analogy cuts both ways. The developer created the tool, not the specific artwork. A camera maker does not own the photographs their cameras produce, and by the same logic, the developer's role in any particular AI-generated image is indirect. The developer could plausibly be considered an author only in the narrower case where they directly create a work using the AI themselves — but in the ordinary case of a third party prompting a finished system, the developer's contribution is to the instrument, not to the work.
The AI as Author: A Legal Impossibility (Currently)
Could the AI model itself be the author? The Houston Law Review analysis considers arguments advanced by scholars such as Ryan Abbott, who have pressed for recognizing machine creators. But Caldwell's article counters that the AI "would not be the inventor of a patent due to an incapability to hold the rights of said patent" — a point reinforced by the Naruto v. Slater litigation, in which a non-human claimant's authorship rights were rejected. U.S. copyright law currently requires human authorship, which makes the AI model categorically ineligible to hold copyright, regardless of how sophisticated its outputs become.
The User as Author: Navigating Human Contribution
The end user is where the law bends most. The user initiates the artwork's existence and may select, arrange, or modify AI-generated material. The Copyright Office's position is that registration depends on "the extent to which the AI tool is part of the creative process" — the more human creativity involved, the more likely a work can be registered. A Georgetown University library guide summarizing the current situation puts it directly: AI-generated art and text "is not copyrightable on its own," but human contribution can change that outcome.
The practical boundary is the prompt. A prompt alone, in the Office's view, leaves the traditional elements of authorship to the machine. But when a human selects, arranges, or modifies AI-generated material in a sufficiently creative way, the AI becomes merely an assisting instrument, and the human's contribution may qualify for protection. The extent of human creative control is the deciding factor.
Who Can Claim Authorship? A Framework for AI Art
The table below applies the competing legal and philosophical theories of authorship to the three entities involved in AI art creation, showing why U.S. law struggles to assign copyright and defaults to a human authorship requirement.
| Potential Author | Argument for Authorship | Argument Against Authorship | Current U.S. Legal Status |
|---|---|---|---|
| AI Developer | Created the system that enables the art's existence, analogous to a camera manufacturer enabling a photograph. | Built the tool, not the specific artwork; the developer's role is indirect unless they directly create a work using the AI. | No authorship claim to works generated by third-party users of their tool. |
| AI Model | Could be seen as producing the "traditional elements of authorship" in response to a prompt. | As a non-human entity, it cannot hold copyright; U.S. law requires human authorship, as the Naruto v. Slater litigation reflects. | Ineligible to hold copyright; works produced solely by a machine, even with a human prompt, lack human authorship. |
| End User | Sets the artwork's existence in motion and provides creative input through prompts, selections, or modifications. | A prompt alone leaves the traditional elements of authorship to the machine, per the Copyright Office's 2023 guidance. | Copyright possible only if human contribution meets the originality requirement, with the AI as an assisting instrument. |
The framework's result is clear: two of the three candidates fail on current law, and the third succeeds only conditionally. That is precisely why the human authorship requirement has become the default — and why the burden of proof falls on the human who wants protection.
Economic and Ethical Challenges for Human Artists
The stakes extend well beyond legal theory. Artists whose copyrighted work is used in AI training datasets report a specific harm: the AI can later produce art in their style, which these artists say takes away the profitability of their original work. Caldwell's Houston Law Review analysis notes their argument that extending copyright protection to AI art would make the practice more ethical, because developers would have reason to be more conscientious about the images used in training, restoring creative integrity and market share the artists claim has been taken from them.
Policy commentary from Brookings identifies a second, subtler disruption: because the cost of creating AI-generated art is near zero, artists may be incentivized to create noncopyrightable works for potentially large financial payoffs. That inverts the traditional bargain of copyright, which exists to safeguard artistic labor. If financial success no longer depends on securing copyright protection, human "original authorship" may lose its value as AI-generated content becomes more cost-effective — throwing the foundational purpose of copyright into question. Ongoing lawsuits, artist reactions, and market shifts, Brookings observes, all highlight the struggle to define human authorship and protect artists' rights when AI-generated works carry significant commercial value but lack clear copyright protection.
Navigating the Future of AI Art and Human Creativity
For artists and creators engaging with AI tools, the practical decision is this: understand the U.S. Copyright Office's human authorship requirement, and actively ensure your creative input — selection, arrangement, modification — is substantial enough to qualify for protection, rather than relying on a prompt alone. This article addresses U.S. copyright law only, does not provide legal advice, does not predict the outcome of pending lawsuits, and does not cover patent or other intellectual property regimes.
The measurable indicator to watch is the U.S. Copyright Office's future guidance or court rulings on specific AI-assisted works — particularly anything that clarifies the threshold for "sufficient human authorship" or addresses compensation for the use of artists' work in training data. The Office has already said it is monitoring developments and may issue additional guidance; the next such ruling will move the boundary that every AI artist, developer, and human illustrator currently works around.
