On September 19, 2024, the US Court of Appeals for the DC Circuit heard arguments in the matter of Thaler v. Perlmutter on the question of whether an image “autonomously” generated by artificial intelligence can be copyrighted. The appeal followed a decision by the DC District Court (687 F.Supp.3d 140) finding that copyright law requires human authorship of a work to qualify for copyright protection. As part of his copyright application, Appellant Thaler confirmed that the image in question was “autonomously generated by an AI” and “lack[ed] human authorship." Based, in part, upon the contents of Thaler’s application the district court ultimately found that: “In the absence of any human involvement in the creation of the work, the clear and straightforward answer is the one given by the Register: No.
The District of DC Appellate panel (the panel) seemed unconvinced by Mr. Thaler’s arguments. The panel noted that it appeared that Thaler always intended to set this matter up as a test case and it was only on appeal that he suggested that he provided the requisite human input by programming the machine and algorithm that created the image. The panel seemed disinclined to consider Mr. Thaler’s new arguments that were not part of the administrative record. Based on the tenor of the oral arguments and the questions by the panel, we believe it likely that the DC Circuit will uphold the district court decision and find that copyright law requires some level of human input, while leaving open the question of exactly how much input will be enough to obtain copyright protection.
Finally, we note that the Copyright Office has already identified factors it will consider in determining if an application meets the "human requirement," including:
- How did the user define the prompt to AI? The level of detail or instruction as to tone or style may influence whether the author is a person.
- How did the user respond to what AI generated? A user sending the design back multiple times or making detailed modifications could demonstrate the alignment of the user and machine with the user’s vision.
