Federal Court Confirms That Copyright Registration Requires Some Human Input
A recent ruling by the D.C. Federal District Court ’s in the Thaler case affirmes the Copyright Office’s refusal to issue a registration for a work generated solely by Artificial Intelligence (“AI”).
Stephen Thaler, the owner of a computer system that he used to create a generative digital visual artwork entitled, “A Recent Entrance to Paradise”, attempted to register the work with the Copyright Office, naming the machine as author and claiming it was “autonomously created by a computer algorithm running on a machine.” Thaler v. Perlmutter, et. al., 1:22-cv-01564, 3 (D.D.C. 2023).
The Copyright Office denied the application on the basis that the work lacked the human authorship necessary to support a copyright claim. Thaler requested reconsideration from the Copyright Office twice and lost in each instance. In what is clearly a test case for AI authorship, he filed a challenge with the federal court arguing the Copyright Office’s denial was arbitrary, capricious, and an abuse of discretion.
In an easy case for dismissal, the federal court stood by the Copyright Office and agreed that human authorship is the bedrock requirement of copyright. While humans can use machines as tools in the creation of work, such as a camera, where the claimant concedes that the work was created autonomously by machine, the work lacks any human authorship.
This case is consistent with the recent guidelines published by the Copyright Office in registering works that may be generated using artificial intelligence tools https://copyright.gov/ai/ai_policy_guidance.pdf. If works are generated using more than a minimal amount of AI, that use of AI must be disclosed and the portion created using AI disclaimed, with the Office registering the aspects of the work claimed on the application as human authored. In other words, merely creating a work generated by text prompts, without anything more may not be registerable. The Office needs to see human authorship associated with the work seeking registration.
While the court’s decision in Thaler was not surprising as no human authorship was claimed , a more interesting case will be where an application for a work that was created by a human using a generative AI computer system, and by continually using prompts to modify and manipulate the results creates a desired work. Perhaps, then there would be enough of a human hand to guide the results to satisfy the requirements of authorship. Thaler, by identifying only the machine as author in the application, seemed to be begging the Office to refuse the application in order to question the understanding of what an author is under the Copyright Act. The court in this instance had no problem finding the author must be human and not eligible for registration.
The Copyright office has been actively studying AI with a number of listening sessions held this past year and has a page devoted to its AI initiatives. https://copyright.gov/ai/ummer.
Its most recent initiative from the Copyright Office is the Notice of Inquiry (“NOI”) published August 30, 2023 seeking comments for a study of the copyright law and policy issues raised by artificial intelligence systems. https://www.govinfo.gov/content/pkg/FR-2023-08-30/pdf/2023-18624.pdf The Copyright Office is looking for comments from a wide variety of the public, including those involved in the use of copyrighted works to train AI models, the appropriate levels of transparency and disclosure with respect to the use of copyrighted works, and the legal status of AI generated outputs. Comments are due Wednesday October 18, 2023. While the General questions cover many topics, responders need only answer those questions which are applicable to their knowledge or interest. DMLA legal committee will be working on a response to file on behalf of DMLA, but all members are encouraged to file individual responses to questions where they have knowledge of the issues raised.