AI and Copyright: Can Art Created by Generative AI Qualify for Copyright Protection?
Computer Scientist Stephen Thaler aspires to expand the definition of authorship in the copyright registration context. If his appeal is dismissed, this will affirm that AI generated works of art do not satisfy the requirement of traditional-human authorship in an application for copyright protection.
Should the creative outputs of generative AI systems be eligible for copyright protection? Stephen Thaler thinks so. Thaler recently filed an appeal after the U.S. Copyright Office (the “Office”) rejected his copyright application for a work of art made by his generative artificial intelligence (“AI”) system.
Stephen Thaler creates and owns AI programs that create works of art. Thaler alleges his system—coined the Creativity Machine—produces art pieces synonymous with those made by a human artist. Thaler ran an algorithm on the Creativity Machine that prompted the system to autonomously compose the 2-D work of art detailed in his copyright application.
The Office denied Thaler’s application once and his reconsideration requests twice, contending that human authorship is a prerequisite for registering a copyright. The Office wrote in their various refusal letters that the Creativity Machine’s work lacked that necessary element of human creation. In the copyright registration context, the Office’s Compendium defines a claimant as “either the author of the work that has been submitted for registration or a person or organization that owns all of the rights under copyright that initially belonged to the author of that work.”
Therefore, to be a valid claimant, Thaler must prove the artwork is eligible for copyright protection, and that he either authored or owns the rights to the work. Thaler contends that he is permitted to register the piece as a work-made-for-hire and, therefore, the copyright was transferred to him even though he did not author the piece.
A work-made-for-hire has two elements: (1) the work must be “prepared by an employee within the scope of their employment” and (2) “specially commissioned for use as a contribution to a collective work . . . .” In a responding motion, the Office states that the doctrine of work-made-for-hire must deal with an employee or agent of the claimant that has created the work and signed off the rights to that work to the claimant. The implications of the court ruling for Thaler’s use of the doctrine would legally exempt copyright registration applications from an element of human authorship; this directly contradicts the federal policy of the Office and policy initiatives behind The Copyright Act.
On August 18, 2023, the U.S. District Court of D.C., sided with the Office and ruled against Thaler. Judge Howell was unconvinced by Thaler’s arguments, and held that the Copyright Act’s authorship requirement is “presumptively human” and necessary regardless of whatever programmed instructions and direction Thaler gave to his AI.
The Copyright Act is the foundation of modern copyright law; it details a broad overview of rights for copyright holders. Thaler’s suit argues that the framers of the Act failed to detail what authorship entailed. He relies upon section 102, which states that “copyright protection subsists . . . in original works of authorship fixed in any tangible medium of expression, now known or later developed, from which they can be perceived, reproduced, or otherwise communicated . . . .” Thaler’s view of section 102 is that human creation of the work is not a requisite element detailed by the act necessary to approve a copyright registration.
The court affirmed the Office’s policy, that without human involvement, an autonomously generated creation of a computer system is not eligible for copyright protection. The court found Thaler’s contention that AI’s authorship and copyright should transfer to him via the “work-made-for-hire doctrine” as irrelevant, as there was no copyright protection present in the first place that Thaler could lay a claim over or transfer.
The Judge also noted that while U.S. Copyright law was historically adapted to be malleable, integrating modern technologies in the creative process and end product, the Copyright Act and its significance today are to protect creative incentive by protecting their original authorship.
Thaler’s case is a harbinger of further AI-situated court battles. His case raises inquiries on the protection of human ingenuity as it relates to artistry and the rise of AI. The courts are likely to face challenging considerations on how much human input is necessary to qualify a user or creator of an AI system as an author of a generated work. Furthermore, it raises the question of whether creating an AI system that can generate a work of art, or imputing the algorithm that compels the system to make a piece is evidence of human authorship.
Currently, the understanding of AI’s role in art is that to apply for a copyright, the part of AI involvement in artistry must be trivial. Used in Thaler, this standard of trivial creation would not apply to the Creativity Machine’s product as it was work solely generated by an AI and lacks any sort of human authorship. It will be interesting to see how the court will analyze Thaler’s appeal as he argues that the Copyright Act does not explicitly limit copyright protection to works made by human authorship, and the authorship standard is satisfied by the work made for hire doctrine.
As AI’s role in the creative process becomes more common place, the court’s adjudication of rejected Copyright claims—such as Thaler—will help creators navigate their usage of AI when they create art they seek to copyright in the future.