BitLaw

No Copyright Protection for Works Generated Solely by AI

The U.S. Copyright Office and courts have consistently ruled that works created without human authorship are ineligible for copyright protection. This principle has been tested in a variety of legal contexts, including the well-known "Monkey Selfie" case and more recent disputes concerning AI-generated content.

One of the most famous cases involving non-human authorship is Naruto v. Slater, which stemmed from a series of photographs taken by a macaque monkey using a camera belonging to British wildlife photographer David Slater. Slater had set up his camera in an Indonesian jungle and, in an effort to capture unique images, allowed the monkeys to interact with the equipment. One particularly curious monkey, later named Naruto, managed to press the shutter button, resulting in a set of striking self-portraits (including the one below) that sparked an international copyright dispute.

The case arose when the animal rights organization People for the Ethical Treatment of Animals (PETA) filed a lawsuit on behalf of Naruto, the macaque monkey, arguing that he should be recognized as the copyright holder of the images. PETA contended that since Naruto had independently taken the photographs by pressing the shutter button, he should be afforded the same intellectual property rights as a human photographer. The lawsuit was filed against David Slater, the owner of the camera, who claimed copyright over the images, asserting that his creative input in setting up the equipment and enabling the monkey’s interaction justified his authorship. The U.S. District Court dismissed the case, ruling that copyright law does not protect creations made by non-human entities, a decision later upheld by the Ninth Circuit Court of Appeals on the ground that animals lack standing under the Copyright Act.

This issue came to the forefront again in Thaler v. Perlmutter, where computer scientist Stephen Thaler sought copyright protection for an image generated entirely by his AI system (shown to the right). Thaler, a vocal advocate for AI’s recognition in intellectual property law, was also behind a similar legal dispute with the U.S. Patent and Trademark Office, where he argued that an AI system should be recognized as an inventor on a patent application. His efforts to challenge existing legal frameworks in both copyright and patent law have sparked significant debate about the role of artificial intelligence in creative and inventive processes.

Thaler claims that the work at issue, titled "A Recent Entrance to Paradise," was generated by Thaler’s AI system without human intervention. The U.S. Copyright Office rejected Thaler's application for copyright registration on behalf of his AI software, reasoning that human involvement in the creative process is a fundamental requirement for copyright eligibility. Thaler challenged this decision, arguing that the law should evolve to accommodate AI-generated creativity. However, the U.S. District Court for the District of Columbia reaffirmed the Copyright Office’s position after Thaler appealed the initial rejection. The court ruled against Thaler, emphasizing that copyright law has consistently required human authorship. Thaler then appealed to the U.S. Court of Appeals for the D.C. Circuit, which upheld the lower court’s ruling, reinforcing that works created solely by AI do not qualify for copyright protection under existing statutes.

Copyrightability Report of January 2025--Separability

In its January 2025 Copyright and Artificial Intelligence, Part 2: Copyrightability report, the U.S. Copyright Office confirmed its prior stance that a work must have a human author to be eligible for copyright protection. While AI can function as a tool to aid in the creative process, it cannot serve as the sole creator of a copyrightable work. This distinction aligns with the long-established principle that copyright law protects only works that exhibit human intellectual effort and expressive choice. The report reaffirmed that merely using AI to generate content—even with extensive prompting, iteration, and refinement—does not satisfy the human authorship requirement.

The report drew parallels to previous copyright cases, including Community for Creative Non-Violence v. Reid (CCNV), which clarified that authorship entails original expression created through human decision-making. The Copyright Office further emphasized that prompts, regardless of their complexity or specificity, do not amount to authorship. Even when users iteratively refine prompts to direct AI output, the process remains distinct from human authorship because the AI system independently determines the final expressive elements. This conclusion is consistent with the Feist Publications, Inc. v. Rural Telephone Service Co. decision, which established that copyright protection cannot be based on mere effort or investment—often referred to as the “sweat of the brow” doctrine.

The report also highlighted how courts differentiate between AI as a creative tool versus AI as the primary source of expression. When an AI system generates content independently, without meaningful human input, the resulting work is not eligible for copyright. However, if a human contributor meaningfully shapes, arranges, or modifies AI-generated material in a way that reflects original human authorship, copyright protection may be granted for those specific human contributions.

The Copyright Office pointed to this example of an artist who submitted a hand-drawn image of a cyborg and then used AI tools to refine and enhance the illustration. The AI system processed the original drawing and applied modifications based on textual prompts, altering aspects such as shading, texture, and background details. Despite these AI-generated refinements, the Copyright Office determined that the human-authored portions remained identifiable and distinct within the final work. The registration was granted only for the aspects of the image that could be clearly traced to the artist’s original contribution.

Crucially, the applicant had to disclaim any non-human expression in the final work, meaning that any portions generated purely by AI—such as newly introduced textures, lighting effects, or other modifications that were not directly authored by the human creator—were excluded from the copyright claim. The Copyright Office explicitly stated that registration was limited to the unaltered human authorship that was perceptible and separable from the AI-generated elements.

This "separable" distinction is key to understanding copyright protection for AI-assisted works. If the AI-generated enhancements are so integrated that the original human authorship can no longer be identified as distinct from the machine-generated content, the work as a whole will not qualify for copyright protection. The ability to separate human-created elements from AI-generated modifications remains a fundamental requirement for copyright eligibility in cases involving AI-assisted creativity.

In contrast, works where human and AI contributions are inseparable or where AI-generated content dominates without a clear human imprint remain uncopyrightable. The Copyright Office illustrated the concept of inseparable AI-generated works with the example of Sahni, an applicant who attempted to register an image generated by AI in the style of Vincent van Gogh’s Starry Night. Sahni provided an initial human-created input, yet the AI’s transformation was so pervasive that the human contribution was no longer distinguishable. This is seen in the shown series of images. The AI system applied stylistic modifications that fully replaced or subsumed any discernible original elements, making it impossible to separate human authorship from the machine’s influence. Despite the fact that the original input may have contained copyrightable expression, the final work was deemed uncopyrightable because the human-authored portion could not be isolated from the AI-generated alterations. The Copyright Office determined that the AI system—not Sahni—was the entity responsible for the expressive aspects of the final output, meaning that no human author could be credited. Because copyright protection extends only to human-authored works, and there was no separable human contribution remaining in the final product, registration was denied.

The inability to separate human-created elements from AI-generated expression remains the defining limitation in these cases. When AI is used to refine or enhance human creativity while preserving identifiable human authorship, copyright protection may still be granted. In the diagram shown to the left, a human author collaborated with an AI "author" (not a real author in the copyright sense because non-humans can never be a copyright author). The combined work is as shown on the left hand image, with the AI contribution shown in blue and the human contribution in green. The protection granted for this type of collaboration is very limited, however, because all of the contributions created by the AI must be fully disclaimed. Only the original human contribution is protected in the final work. And this protection is granted only to the extent that the human authorship is identifiable and separable in the combined work. This is the result that the Copyright Office described in connection with the cyborg image shown above.

Compare this result with the result that the Copyright Office reached in Sahni's Starry Night image. In this example, the contributions of the AI and the human are not separable from each other. Conceptually, this is shown in the diagram to the right, with the left hand image showing the blue (AI) and green (human) contributions mixed throughout the entire work. In this scenario, there is no separable human authorship in the final work. Based on the current position of the Copyright Office, no copyright protection is provided for any part of this work.

Joint Authorship in AI-Assisted Works

Joint authorship in copyright law applies when two or more human contributors collaborate with the intent of creating a unified work, each contributing protectable expression. The intention to combine the work must exist at the time the contribution is made. It is not necessary that the contributions be equal in effort or value. If a joint work exists, then both authors are co-owners of an undivided interest in the entire work. More information on joint authorship can be found on Bitlaw's Copyright Ownership page. In its 2025 Copyrightability Report, the Copyright Office has firmly rejected the notion that humans and AI can work together as joint authors. The report clarified that authorship must involve human intellectual input and expressive choice, which AI systems, as mere computational tools, do not possess. A human can never be a joint author with AI, because a non-human entity can never be an author. Thus, the concept of joint authorship was rejected as a framework for analyzing copyright protection for works of authorship created, in part, through artificial intelligence.

The U.S. Patent and Trademark Office has taken a different approach, and has used the concept of co-inventors to help determine when AI-assisted inventions are protected by U.S. patent laws. The full analysis conducted by the Patent Office is described in detail on Bitlaw's AI Inventors and Patent Applications page. Under the PTO's approach, AI will never be identified as an inventor in a patent application because only humans can be inventors. Thus, if AI conceived of a new invention without any human contribution, the invention is not protectable through the patent system because there was no human inventor. Nonetheless, if at least one human contributor has made a meaningful contribution to the invention, the invention is eligible for patent protection. The fact that artificial intelligence participated in the inventive process does not disqualify the invention from protection. All that is required is that a human conceived of a significant portion of the invention such that the human would qualify as a co-inventor had the other party been human and not artificial intelligence. As depicted in the diagram to the left, the protection currently provided under patent law covers the entire invention, even though a portion of the invention may have been conceived by artificial intelligence. All that is necessary is for a human to have been the inventor of a significant aspect of the invention.

The approach to this issue by the Patent Office appears to be a more practical solution for handling joint human/AI activity. The problem with the Copyright Office's current approach is that AI-assisted creative content is rarely produced in a way that allows for clear separation of human and machine contributions. The process of generating AI-assisted works—whether texts, images, or other media—typically involves hundreds of iterative interactions between human creators and AI tools, making it nearly impossible to pinpoint discrete portions that are wholly human-created. This challenge is especially pressing in fields like literature, photography, filmmaking, and music, where AI is becoming an essential creative tool. If the Copyright Office maintains its stance, many of these works will remain unprotected because the requirement for separable human authorship renders them ineligible. This raises significant concerns for creators and industries that rely on AI, potentially leaving vast amounts of human-created, but AI-assisted content outside the bounds of copyright protection.

The most viable solution to this dilemma may be redefining copyrightability in this context in line with joint authorship principles. While AI cannot legally be an author, recognizing the human contributor as the sole legal author of an AI-assisted work—provided that the human has made a substantial and creative contribution—could align with the way copyright law traditionally acknowledges joint authorship among human collaborators. If a human’s role in directing, shaping, and refining the work is equivalent to what would traditionally qualify for joint authorship, it stands to reason that they should be recognized as the work’s author, even if their collaborator is an AI system rather than another human.

Artificial Intelligence (AI) Copyright Attorney

Please see Dan Tysver's bio and contact information if you need any AI-related legal assistance. Dan is a Minnesota-based attorney providing AI advice on intellectual property and litigation issues to clients across the country.