On 18 August 2023, the United States District Court for the District of Columbia ruled that the Copyright Office had not erred when it refused to register copyright protection for a work created by artificial intelligence.
The application to the Copyright Office was filed by Stephen Thaler, the owner of the autonomous system “Creativity Machine,” in which he generated a piece of virtual art titled “A Recent Entrance to Paradise.” The work is meant to depict a near-death experience.
What was it about?
In the registration form, Thaler named the AI (The Creativity Machine) as the author and explained that the work had been autonomously created by the system's algorithm. Thaler sought to have the copyright registered in his favor as the owner of the autonomous system, on the basis of the “work-for-hire” concept (note: under this concept, the employer rather than the creator/employee is regarded as the statutory author).
During the court proceedings, Thaler attempted to argue that he gave the AI direct instructions, that he had the entire process of creating the work fully under his control, and that the AI merely acted according to his instructions. The court, however, declined to consider these points, as they fell outside the scope of review. Thaler had not raised these (subsequently alleged) facts during the registration proceedings. These new assertions were inconsistent with the earlier proceedings.
In the earlier registration proceedings, Thaler had informed the Copyright Office that the work was created independently by the AI system. The Copyright Office decision under review is therefore based on the assumption that Thaler played no role in its creation.
How did the court rule?
Judge Beryl A. Howell acknowledged that, as artists increasingly incorporate AI into their tools when creating new art, new boundaries will need to be drawn in the field of copyright. The growing reduction of human creativity in generating works will raise challenging questions, in particular:
- How much human input is necessary for the user of an AI system to qualify as the author of the generated work?
- What will the scope of protection of generated works be?
- How should the originality of AI-generated works be assessed where the system was trained on unknown pre-existing works?
- How can copyright best be applied so as to encourage creative activity when using AI?
Within the dispute, however, the court could address only a single question: “Whether a work generated by an AI system can be recognized as a copyrighted work.” The court concluded that the answer was NO. There is a lack of human involvement in the creation of the work. A human author is a necessary condition for the valid creation of copyright. In the reasoning of its decision, the court also referred to the previous case law of the U.S. Supreme Court. In the Sarony case, it explained that it is the photographer, as a human author, and not the camera, who conceives and designs the photograph and then uses the camera to capture it. The Naruto v. Slater case, in turn, establishes that a macaque cannot sue for infringement of its copyright.
Although we will have to wait a little longer for the important answers in the field of copyright, the court reached the already anticipated conclusion that an AI system alone cannot be regarded as a human author. The same conclusion is inevitable in the Czech Republic as well, since under the Copyright Act only a human being can be an author. You can read about the first Czech court ruling on this matter here.
Do you want to use AI systems while still being the author? For our clients we provide an AI compliance service, within which we address this question, and entrepreneurs can also take advantage of the subsidized IP scan service. And if you want practical training on where the line lies between authorship and the use of AI systems, come to one of our courses.