Introduction
The US Copyright Office released Part 2 of the Copyrightability Report on January 29, 2025, as part of a three-part series analyzing the economic implications of artificial intelligence (AI) on copyright policy [5]. This report [1] [2] [3] [4] [5] [6] [7], developed by an ad hoc committee of economic scholars [5], addresses the intersection of AI and copyright [5] [6], focusing on the copyrightability of works created using generative AI tools. It emphasizes the necessity of human authorship for copyright protection and discusses the implications of AI-generated content on existing copyright laws.
Description
On January 29, 2025 [1] [2] [4] [6], the US Copyright Office published Part 2 of the Copyrightability Report, a comprehensive analysis that is part of a three-part series examining the economic implications of artificial intelligence (AI) for copyright policy. This initiative [1] [6], which began in early 2023 [6], aims to provide guidance on registering works that incorporate AI-generated material [1] [6]. Developed by an ad hoc committee of economic scholars [5], the report addresses key legal and policy issues at the intersection of AI and copyright, particularly focusing on the copyrightability of works created using generative AI tools. It emphasizes the necessity of human authorship for copyright protection and mandates the disclosure of AI-generated material in registration applications [6].
The report expresses skepticism regarding the copyrightability of outputs generated solely by AI [4], citing insufficient human expressive control as a primary concern. For a work to be considered copyrightable [4], it must be an “original work of authorship,” which has been interpreted to require significant human involvement in the creative process [4]. It discusses the long-term social welfare gains from creative works and the impact of copyright policy on innovation and cultural development [5], especially in relation to generative AI technologies, such as large language models [5]. The report clarifies that purely AI-generated outputs are not protected under current copyright law due to the lack of necessary human control over expressive elements. However, if a human inputs original works as prompts—such as an original drawing—and that work is perceptible in the AI-generated output [2], copyright protection may be available for that portion [2].
The report asserts that existing copyright principles apply to generative AI without necessitating legal changes [3] [7], although these principles offer limited protection for many AI-generated works [3] [7]. It distinguishes between human-generated and AI-generated works [5], noting that AI systems cannot hold copyrights and that the nature of user prompts—whether simple or complex—does not establish authorship [7]. The report explicitly rejects the notion that merely creating text prompts for AI qualifies a human as the author of the outputs [1], emphasizing that the AI’s interpretation of user directions leads to varied results. It also discusses various use cases of generative AI and their implications for the copyrightability of creative works [4], clarifying that using generative AI for ideation or brainstorming does not inherently affect the copyrightability of a final work [4]. The distinction is made between using generative AI as a tool for human creativity versus using it as a substitute for human authorship [4] [7].
To illustrate the difference between AI-generated works and human authorship, the report references Jackson Pollock’s painting technique [3], emphasizing that Pollock’s creative authority over key artistic choices demonstrates a level of human control that is absent in AI-generated content [3]. While AI-generated elements are generally uncopyrightable [3] [7], the report notes that some protection may apply when artists modify their own work using AI, provided the original work remains recognizable [3] [7]. For instance, an artist enhancing an illustration with AI-generated effects may retain copyright protection if the original work is still evident [3]. Works that incorporate AI-generated content may also be eligible for copyright if they involve significant human creative input [3] [7], such as a comic book featuring AI-generated images arranged by a human or a film with AI-generated special effects.
The report examines the eligibility of AI-generated content for copyright protection and the potential displacement of demand for human-generated works [5]. It addresses the risks of AI systems producing outputs that may infringe upon existing copyrighted materials and the challenges of determining liability in such cases [5]. Additionally, it discusses the commercial exploitation of personal identifiers by AI and the implications for rights of publicity and individual control over personal identity [5]. The report investigates the effects of using copyrighted works as training data on creator incentives and the overall quality and quantity of creative output [5].
Furthermore, the report highlights the challenges AI developers face in accessing large datasets that include copyrighted material and discusses potential policy measures for regulating the inclusion of such works in AI training datasets [5]. It considers negotiation mechanisms and safeguards to manage rights associated with training data [5]. The report emphasizes the critical issue of human control over the creative process [7], advocating for equitable policy design that considers diverse stakeholder impacts [5].
Lastly, it underscores the need for updated economic analyses and policy frameworks in response to technology-driven shifts and calls for further research to understand the economic impacts of AI on creative industries [5], while also examining the broader socioeconomic implications and potential biases in AI-related copyright policies [5]. The report concludes that no new laws are needed for questions of copyrightability in AI [1], relying instead on established copyright principles regarding human authorship and originality [1]. It acknowledges that advancements in technology may alter these conclusions in the future [1]. Following the Loper Bright decision [2], the report notes that while influential, it does not have final authority on copyright matters; courts will ultimately determine the protectability of AI-generated outputs [2], with the possibility of escalation to the Supreme Court if necessary [2]. The anticipated third report is expected to cover the training of AI models on copyrighted works [2], licensing considerations [2] [6], and liability allocation [2], which could significantly influence the AI market and its future [2].
Conclusion
The Copyrightability Report underscores the critical role of human authorship in copyright protection, particularly in the context of AI-generated content. It highlights the challenges and implications of integrating AI into the creative process, emphasizing the need for existing copyright principles to adapt to technological advancements. The report calls for further research and updated policy frameworks to address the economic and social impacts of AI on creative industries, while acknowledging that future technological developments may necessitate reevaluation of these conclusions.
References
[1] https://www.jdsupra.com/legalnews/copyright-office-issues-report-on-2232930/
[2] https://natlawreview.com/article/copyright-office-copyrighting-ai-generated-works-requires-sufficient-human-control
[3] https://btlaw.com/insights/alerts/2025/copyright-office-says-ai-generated-works-based-on-text-prompts-are-not-protected
[4] https://www.lowenstein.com/news-insights/publications/client-alerts/us-copyright-office-issues-report-on-copyrightability-of-generative-ai-outputs-ai
[5] https://chatgptiseatingtheworld.com/2025/02/12/us-copyright-office-issues-report-on-economic-implications-of-ai-for-copyright-policy/
[6] https://www.wiley.law/alert-Generative-AI-in-Focus-Copyright-Offices-Latest-Report
[7] https://natlawreview.com/article/copyright-office-says-ai-generated-works-based-text-prompts-are-not-protected
