Introduction

Artificial Intelligence (AI) is everywhere and is increasingly doing many things that humans can do. AIs can paint, sing, dance, and write. Can’t sing? Not a problem! You can generate AI vocals with human-like tone and pitch using a growing number of AI tools such as Kits AI, ACE Studio, Synthesizer V, Lalals, and Voicemod. The Kits AI Singing Voice Generator is described as “a cutting-edge tool designed to create vocals, melodies, and harmonies using artificial intelligence.” Can’t paint? No worries! Turn to a text-to-painting AI generator like DALL·E 2, Midjourney, OpenArt, and NightCafe. Wonderful, right? But, does US copyright law protect works created entirely by an AI? Recently, the United States Court of Appeals for the District of Columbia Circuit (DC Circuit) answered this question in the case Thaler v. Perlmutter. On March 18, 2025, the DC Circuit ruled that an AI cannot be an author under the US Copyright Act of 1976 (Copyright Act). In a unanimous opinion delivered by Circuit Judge Miller, the court concluded that the Creativity Machine, the AI in issue in the case, “cannot be the recognized author of a copyrighted work because the Copyright Act … requires all eligible work to be authored in the first instance by a human being” (emphasis added).

Facts of the Case

Dr. Stephen Thaler, a computer scientist, created a generative AI named the “Creativity Machine.” The Creativity Machine produced a picture that Thaler titled “A Recent Entrance to Paradise” (the Work). Thaler subsequently submitted a copyright registration application for the Work to the US Copyright Office, the federal agency that administers the Copyright Act. On the application, Thaler listed the Creativity Machine as the sole author of the Work and listed himself not as a co-author but as the Work’s owner. Also on the application, Thaler left a note for the Copyright Office stating that the Work “was autonomously created by a computer algorithm running on a machine” and that he was “seeking to register this computer-generated work as a work-for-hire to the owner of the Creativity Machine.”

In August 2019, a Copyright Office registration specialist refused to register the Work finding that it “lack[ed] the human authorship necessary to support a copyright claim.” Thaler made a first request for reconsideration pursuant to 37 CFR § 202.5(b). On March 30, 2020, the Copyright Office’s Registration Program refused to register the Work finding that it “lacked the required human authorship necessary to sustain a claim in copyright” and noting that Thaler had “provided no evidence on sufficient creative input or intervention by a human author in the Work.” Thaler filed a second request for reconsideration pursuant to 37 CFR § 202.5(c). On February 14, 2022, the Copyright Office’s Review Board denied Thaler’s request holding that under the human-authorship requirement, to be eligible for copyright registration, a work must be authored in the first instance by a human being. Thaler sought judicial review in federal district court, arguing that the human authorship requirement was unconstitutional and unsupported by statute or case law. In 2023, Judge Beryl A. Howell of the US District Court for the District of Columbia affirmed the Copyright Office’s decision concluding that “[h]uman authorship is a bedrock requirement of copyright.” Thaler appealed to the DC Circuit.

Fig. 1. A Recent Entrance to ParadiseSource: As cited in the Copyright Review Board’s Decision

The Decision of the DC Circuit

The DC Circuit unanimously affirmed the district court’s decision denying copyright protection to the Work stating that “[a]s a matter of statutory law, the Copyright Act requires all work to be authored in the first instance by a human being.” Writing for the court, Judge Millett started by noting that “[a]uthors are at the center of the Copyright Act.” Section 102(a) of the Copyright Act declares that copyright protection extends to “original works of authorship” that are fixed in a tangible medium of expression. But who is an “author”? The Copyright Act does not define the term “author.” Turning to “traditional tools of statutory interpretation,” the court concluded that these tools show that, within the meaning of the Copyright Act, “author” refers only to human beings. Millet pointed to several provisions of the Copyright Act that, together, support the position that an “author” must be human including the provisions relating to ownership (17 USC § 201(a)), inheritance (17 USC § 203(a)(2)(A)), transfers (17 USC § 204(a)), and duration (17 USC § 302(a)). “All of these statutory provisions collectively identify an “author,” and they “make sense only if an author is a human being,” Millett concluded. Ultimately, the court rejected Thaler’s arguments that the term “author” is not confined to human beings.

Is the Decision Consistent With the Decisions of Other Courts of Appeal and the Position of the Copyright Office?

The DC Circuit’s opinion is in line with decisions of other courts of appeal and the position of the Copyright Office. In the case of Urantia Found. v. Kristen Maaherra, the Ninth Circuit held that a book containing words “‘authored’ by non-human spiritual beings” can only gain copyright protection if there is “human selection and arrangement of the revelations.” In a 2018 case, Naruto v. Slater (the “monkey selfie” case), the US Court of Appeals for the Ninth Circuit held that a monkey could not own the copyright in a photograph it snapped of itself.

The DC Circuit’s position is also consistent with the Copyright Office’s longstanding rule requiring a human author. According to the Compendium of US Copyright Office Practices, the Copyright Office “will register an original work of authorship, provided that the work was created by a human being” and “will refuse to register a claim if it determines that a human being did not create the work.” In Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence, released in 2023, the Copyright Office reiterated the human authorship requirement. Finally, in a its recent report: US Copyright Office, Copyright and Artificial Intelligence, Part 2: Copyrightability (2025), the Copyright Office clarified that “if content is entirely generated by AI, it cannot be protected by copyright” because copyright protection is limited to works of “human” authorship.

What Did the Court Not Decide?

First, the Court did not decide whether an AI can or cannot generate quality art. This case was not about the quality of work an AI tool can generate. Over a century ago in Bleistein v. Donaldson Lithographing Company, the Supreme Court established the so-called non-discrimination rule. Essentially, copyright law cannot and does not decide what is and is not “good” art. What copyright law decides is whether the standard for copyright ownership set forth in the Copyright Act has been met and who is the author of a protected work.

Second, the DC Circuit did not decide the status of works consisting of both human-authored and AI-generated materials under the Copyright Act and did not address Thaler’s argument that he was the Work’s author by virtue of making and using the Creativity Machine. The case was strictly about work created entirely by an AI. Are AI-assisted works entitled to protection under the Copyright Act? While works created entirely by AIs do not qualify for copyright protection, it is not clear how much human involvement in the creative process is needed to make a work protectible. The copyrightability of such works would depend on a difficult line-drawing exercise between the respective roles of AI and humans in generating the final output.

Third, the DC Circuit did not address Thaler’s argument that “the Human Authorship Requirement is unconstitutional.” This means that the battle over whether the US Constitution does or does not mandate human authorship will have to be fought at a later date.

Finally, the DC Circuit did not decide anything related to patent law or trademark law. This case was about copyright law. However, for anyone wondering whether an AI can own a patent, wonder no more! In 2022 in Thaler v. Vidal, the US Court of Appeal for the Federal Circuit held that only natural persons can be named on a patent application as inventors or joint inventors. The Supreme Court subsequently declined to grant certiorari in the case.

Key Takeaways and Practical Tips

The decision in Thaler v. Perlmutter has serious implications for businesses and individuals using AI-generated images, whether or not for profit.

• To be eligible for copyright protection under the Copyright Act, the human authorship requirement must be met because copyright law only protects the original expression in a work created by a human author.
• Works created entirely by AI tools are not eligible for copyright protection. As the Copyright Office put it, “[c]opyright does not extend to purely AI-generated material, or material where there is insufficient human control over the expressive elements.”
• Works created by humans with the assistance of AI could be protectible. In other words, “[c]opyright protects the original expression in a work created by a human author, even if the work also includes AI-generated material,” and “[t]he use of AI tools to assist rather than stand in for human creativity does not affect the availability of copyright protection for the output.”
• Individuals and businesses that use AI to generate content must routinely review and monitor their AI-generated work product mindful that such product may not be copyrightable and that whether a work is protectible “may entail establishing the predominance of the human input through a careful record of the use of AI, if any, in the creative process.”
• Care is needed when completing the application for copyright registration. Unless an AI is the actually the sole author of a work, an applicant for copyright registration should refrain from listing the AI as the sole author. Using the Standard Application, clearly identify the author(s) and provide a brief statement in the “Author Created” field that describes the authorship that was contributed by a human. The Copyright Office advice that applicants “should not list an AI technology or the company that provided it as an author or co-author simply because they used it when creating their work.”
• Because the Thaler decision was based on the court’s reading and interpretation of the Copyright Act, the issue of whether AIs qualify as authors under the Copyright Act now reverts to Congress. Whether Congress will choose to expand the definition of authorship and extend copyright protection to purely AI-generated material remains to be seen. However, the Copyright Office takes the position that “[q]uestions of copyrightability and AI can be resolved pursuant to existing law, without the need for legislative change.”

Conclusion

Throughout history, technological developments have challenged the scope and boundaries of intellectual property law. The field of generative AI is raising novel intellectual property rights issues. With the field of AI progressing at lightning speed, interesting legal challenges await in the US and in jurisdictions across the globe. As the use of AI tools becomes more pervasive, more legal challenges around AI systems can be expected.

For humans, the bad news is that AI can and is doing a lot of things that humans can do. The good news is that in the US, works created solely by an AI are not entitled to copyright protection. For now, humans must find consolation in the DC Circuit’s determination that all copyrightable works must “be created in the first instance by a human being, including those who make work for hire.”

Uche Ewelukwa Ofodile (SJD, Harvard) holds the E.J. Ball Endowed Chair at the University of Arkansas School of Law where she has taught a broad range of courses including Public International Law, Intellectual Property Law, Copyright Law, Trademark Law, Technology and the Law, and International Trade and Investment Law. Prof. Ofodile is a life member of the Council on Foreign Relations and was a two-time Senior Fellow of the Mossavar-Rahmani Center for Business and Government at the Kennedy School of Government at Harvard University.