Copyright and Generative AI: A Look at Part 2 of U.S. Copyright Office Artificial Intelligence Report 

Feb 5, 2025 | Northern Virginia Graphics, Inc.

Technological advancements have historically influenced copyright law, from the advent of photography to digital editing. The question now in 2025 is how AI-generated outputs fit within the legal framework. The U.S. Copyright Office launched an AI Initiative and Notice of Inquiry (NOI) in 2023 to address these concerns, gathering input from over 10,000 stakeholders, including creators, publishers, and tech companies. On January 29, the USCO published the second part of a three-part report responding to the public’s input. This second part focuses on copyrightability as it pertains to works made with, or partially made with, generative AI technology.

Under existing U.S. law, copyright protection requires human authorship. The Copyright Office’s report affirms this principle, drawing on key legal precedents such as Feist Publications, Inc. v. Rural Telephone Service Co. (1991), which established that originality is a prerequisite for copyright, Burrow-Giles Lithographic Co. v. Sarony (1884), which determined that mechanical tools do not negate copyright if a human contributes creative elements, and Thaler v. Perlmutter (2023), which held that AI-generated works without direct human input are not eligible for copyright protection. The report categorizes AI involvement in creative processes, distinguishing between AI as an assistive tool and AI-driven generation. It concludes that AI-assisted works may qualify for copyright if human creativity is evident, whereas purely AI-generated content does not meet legal standards. Additionally, the report clarifies that based on the functioning of current generally available technology, prompts alone do not provide sufficient control over the output to meet the standard of human authorship.

Part of this second part of the report is an investigation into how international jurisdictions are approaching the copyrightability of AI-generated works. The USCO looks at the European Union, the United Kingdom, China, Japan, and South Korea, which are also responding to legal issues concerning artificial intelligence and the protection of intellectual property, and it underscores the importance of international legal trends in shaping future policy decisions. The report concludes that despite differing approaches, no country currently grants full copyright protection to purely AI-generated content.

The report examines arguments for and against updating copyright law to address AI-generated content. Proponents of reforming the law argue that extending copyright protection could encourage innovation, strengthen the U.S.’s competitive stance in AI-driven industries, and clarify legal ambiguities surrounding AI’s role in creative processes. Critics caution that AI-generated content could saturate the market, devaluing human creativity, and that the complexity of defining the threshold of human involvement necessary for copyright warrants a case-by-case approach. Ethical and legal risks concerning AI’s reliance on existing copyrighted materials also remain a concern.

In its Part 2 to the Copyright and Artificial Intelligence Report, the Copyright Office concludes that the current technology does not warrant legislative change and that existing legal principles apply to works currently being created by generative AI. As such, existing law determines that purely AI-generated material such as the output of a sole prompt is not copyrightable, and what constitutes something to be purely AI-generated should be addressed on a case-by-case basis. The report also specifies that human authors are entitled to copyright in their works of authorship that are perceptible in AI-generated outputs. This specification may be one upon which we can expect the USCO to expand as it prepares to publish Part 3 of its report, which will pertain to the training of AI Models using existing works.

Source: Copyright and Artificial Intelligence, Part 2 Copyrightability Report