Key Takeaways
- Non-human machines cannot be authors under the Copyright Act of 1976.
- The human authorship requirement does not entirely preclude copyright eligibility for works created with the assistance of AI, provided that there are human-authored contributions to such works.
- In the DC Circuit’s view, adhering to this interpretation of the human authorship requirement does not impede but rather incentivizes humans to create original works in collaboration with AI technologies.
Factual and Procedural Background
On March 18, 2025, the United States Court of Appeals for the District of Columbia held that a non-human machine cannot be an author under the Copyright Act of 1976 (the “Act”). Specifically, the court found that a machine cannot be considered an author of a copyrighted work because the Copyright Act of 1976 “requires all eligible work to be authored in the first instance by a human being.” In a unanimous opinion in Thaler v. Perlmutter, authored by Judge Patricia A. Millett, the appellate court affirmed the District Court’s denial of plaintiff Stephen Thaler’s copyright application.
On May 19, 2019, Thaler applied to register a visual artwork, entitled “A Recent Entrance to Paradise,” created entirely on Thaler’s custom-built, generative AI program, the “Creativity Machine.” In his copyright application, Thaler listed the “Creativity Machine” as the sole author of the artwork and specifically noted that the work was “created autonomously by machine.” The Copyright Office denied registration because “a human being did not create the work.” Thaler sought reconsideration, noting the work was “autonomously generated by an AI” and arguing that the human authorship requirement was unconstitutional. The Copyright Office issued another rejection on the grounds that Thaler’s work “was autonomously created by artificial intelligence without any creative contribution from a human actor.” Thaler sought review in the US District Court for the District of Columbia and lost on summary judgment. He subsequently appealed to the Court of Appeals for the District of Columbia.
History of the Human Authorship Principle
Copyright protection is authorized by the language of the Constitution. Article I grants Congress the power “to promote the Progress of Science and useful Arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries.” Copyright promotes “Progress” by providing authors a “limited duration monopoly” over original works of authorship in order to incentivize the creation and distribution of works of authorship. Under the express language of the statute, copyright protection “subsists in original works of authorship” and “vests initially in the author or authors of the work.
The Copyright Office currently only grants registration “of an original work of authorship, provided that the work was created by a human being.” As far back as 1879, the Supreme Court held that copyright law only protects the “fruits of intellectual labor […] founded in the creative powers of the mind.” In 1884, the Court further held that because copyright law remains limited to “original intellectual conceptions of the author,” the Copyright Office will refuse to register a claim if it determines that a human being did not create the work.
The Copyright Office still largely abides by these principles, as reflected in its registrability decisions to date involving generative AI, as well as its other guidance on this topic. In January 2025, it clarified that “copyright does not extend to purely AI-generated material, or material where there is insufficient human control over the expressive elements.” In the Court of Appeals opinion, Judge Millett affirmed that the Act requires all works seeking copyright protection be authored “in the first instance by a human being.”
The DC Circuit Opinion
Because Thaler listed his generative AI program as the sole author, the court easily affirmed the denial of Thaler’s application. The court found that while the Copyright Act does not expressly define the term “author,” statutory interpretation of numerous provisions of the Act “make sense only if an author is a human being.” The court laid out a seven prong analysis: (1) since copyright ownership relies on an author’s legal capacity to hold property, an entity that cannot own property cannot be an author; (2) the Act premises copyright duration on an ascertainable life, and machines do not have “lives”; (3) machines have no spouses or heirs capable of inheriting copyrights; (4) copyright transfers require signatures, and machines have no legal capacity to provide signatures; (5) machines have no domicile or national identity; (6) machines “lack minds” and the ability to intend to create a joint work; (7) every discussion of “machines” within the Act views machines as tools, not authors.
While none of these points serve as necessary conditions for copyright ownership, the Act overall makes humanity a prerequisite for authorship. The court traced a lineage from the Copyright Office’s mid-1960s report to Congress through to the 1976 passage of the Act to support the conceit that computers were historically never considered capable of acting as authors.
The court quickly dispatched Thaler’s argument that the meaning of “author” is not confined to human beings, noting that the Act would make no sense if “machines” were substituted for authors, as the Act would then, for example, cover a “machine’s children.” Thaler also argued the Act’s Work for Hire provisions allow those who hire creators to be “considered” authors under the Act. However, the court highlighted a critical linguistic difference between considering someone to be an author and defining the hiring party as an author per se.
Finally, Thaler argued that the human-authorship requirement would prevent copyright law from protecting works made with AI. In response, the DC Circuit countered that adherence to the human authorship requirement does not impede copyright protection for several reasons. First, the authorship requirement does not block copyright protection for works made with the assistance of AI. Thaler specifically listed his Creativity Machine as the sole author, obviating any finding of human collaboration. Second, Thaler had not adequately explained how a ban on machines as authors would lead to less original work, given machines do not benefit from economic incentives. Here, the court quickly reaffirmed that copyright law is intended to benefit the public, not the authors. Where Thaler saw the human authorship requirement disincentivizing creativity, the court found the requirement in fact incentivizes individuals like Thaler to create works made in collaboration with AI. Third, Congress’s choice not to amend the Act since 1976, coupled with decades of precedent, all reaffirm the human authorship requirement. Lastly, the court found that the issue of whether the human authorship requirement might eventually impede the creation of original work remained a policy decision best left for Congress.
Despite the clear directives on this issue from the Copyright Office, the federal district court, and the appellate court, this battle may not yet be over: following the release of the opinion, Thaler indicated he plans to appeal the ruling.
