Can a non-human machine be an author under the Copyright Act of 1976? In a precedential opinion issued on March 18, 2025, a D.C. Circuit panel upheld previous decisions from the D.C. District Court and the Copyright Office, ruling that an original artwork solely created by artificial intelligence (AI) is not eligible for copyright registration, as human authorship is required for copyright protection.
Dr. Stephen Thaler developed a generative AI known as DABUS (Device for the Autonomous Bootstrapping of Unified Sentience), also referred to as the “Creativity Machine,” which produced an image titled “A Recent Entrance to Paradise.” In his copyright application to the U.S. Copyright Office, Thaler listed DABUS as the sole author of the artwork, while naming himself as the owner of the work.
In writing for the panel, D.C. Circuit Judge Patricia A. Millett stated that “the Copyright Act requires all work to be authored in the first instance by a human being,” including works created for hire. The court highlighted that the language of the Copyright Act mandates human authorship, as it limits the duration of copyright to the author’s lifespan or a period approximating the length of a human life. The court concluded, “All of these statutory provisions collectively identify an ‘author’ as a human being. Machines do not have property, traditional human lifespans, family members, domiciles, nationalities, mentes reae, or signatures.”
In rejecting Thaler’s claim for copyright on the basis of fully autonomous AI authorship, the court did not address whether Thaler could claim authorship due to his creation and use of DABUS, as Thaler had waived that argument in earlier proceedings. The court also declined to rule on whether an AI-generated work could qualify for copyright protection, but referenced guidance from the Copyright Office, noting that whether AI-generated work is registrable depends on various factors, particularly how the AI tool functions and the extent to which it was used to produce the final work. Recent rulings from the Copyright Office have allowed copyright registration for the human-created portions of “hybrid” AI-human works.
The D.C. Circuit’s analysis aligns with the U.S. patent doctrine, which also requires human inventorship for patent protection, as reflected in the 2022 case Thaler v. Vidal and the USPTO’s Inventorship Guidance for AI-Assisted Inventions issued on February 12, 2024.
The judicial rulings requiring human authorship and inventorship for copyright and patent protection stem from the belief that only humans can “create” art or conceive inventions. This idea underscores the value placed on human creativity, which intellectual property law seeks to protect. This concept was explored in a White Paper published last year by Mammen and a multidisciplinary team of scholars at the University of Oxford. The White Paper identifies three core elements of creativity: (a) an external component (expressed ideas or artifacts that reflect novelty, value, and surprisingness), (b) a mental component (the person’s thought process—combining divergent, convergent, and recognition of relevance thinking), and (c) a social context (society’s recognition of what is new, valuable, and surprising). Intellectual property doctrines require all three of these elements. The paper argues that generative AI does not yet exhibit the mental component essential for human creativity.
Moreover, the White Paper suggests that generative AI could negatively impact human creativity. First, AI-driven creative processes tend to prioritize speed and instant results, encouraging passive consumption over active self-reflection and the back-and-forth thinking that fuels creativity. Second, interaction with AI can erode humans’ confidence in their own creative abilities, leading them to rely on AI-generated “mash-ups” rather than exploring their own creativity.
A study by Doshi and colleagues explored how generative AI impacted short story production when writers used a large language model (LLM) to generate story ideas. The study found that access to AI helped less creative writers produce more enjoyable, well-written stories, but the stories showed less diversity in ideas compared to those written by humans alone. This suggests that while generative AI may augment less creative individuals, it could reduce the overall novelty and diversity of creative works, diminishing collective human creativity.
Dr. Thaler’s arguments challenge the boundaries of existing intellectual property doctrines. Thaler contended that outdated judicial opinions from the Gilded Age cannot settle whether AI-generated works should be copyrighted today. However, as affirmed by the courts and discussed in the White Paper, it is insufficient to argue that AI-generated outputs deserve IP protection simply because they are “just as good as” human-created works. In most instances, AI-created works or inventions involve human input at some stage—whether in creating the AI, setting goals, directing the AI, evaluating the output, or owning the AI. Our DC IP lawyers know that as AI technology evolves, the distinction between human creativity and AI capability may blur, presenting an evolving set of challenges for intellectual property law.