Introduction
The US Copyright Office (USCO) and the US legal system have consistently maintained that human authorship is a prerequisite for copyright protection. This principle has significant implications for works created with artificial intelligence (AI), as such works may not qualify for copyright unless there is demonstrable human involvement. Recent legal cases and guidance have further clarified the boundaries of copyright law concerning AI-generated content.
Description
The US Copyright Office (USCO) may reject copyright applications if the applicant cannot demonstrate human involvement in the creation of the work [5]. Previous guidance indicates that human authorship is essential for copyright ownership [5], and works lacking human authorship do not qualify for copyright protection [2]. In the United States [3], the USCO has established that works created with substantial AI input are not eligible for copyright protection [3], emphasizing that only works attributable to a “human agent” can benefit from such protection [3]. This position has been consistent since 1973 [3], and recent guidance specifies that AI-generated content must be explicitly excluded from registration [3], while human-made elements [3], such as prompt instructions, may be protected [3].
The legal landscape was further clarified by the US Court of Appeals for the District of Columbia Circuit in the case of Thaler v. Perlmutter [1] [4], which ruled that a non-human machine cannot be considered an author under the US Copyright Act [4]. This case arose when Steven Thaler sought to register a copyright for a work created by his “Creativity Machine,” asserting that it was autonomously generated by a computer algorithm [4]. The USCO denied the registration, and Thaler’s lawsuit was unsuccessful [4], reinforcing the established principle that non-human authorship is not recognized under US copyright law [4]. The District Court upheld the USCO’s decision [1], affirming that the art lacked human authorship and thus was not copyrightable [1]. While the ruling does not preclude the possibility of copyrightable works resulting from human-AI collaboration [4], it leaves unresolved questions regarding the rights of AI creators and human users claiming authorship of AI-generated works [4].
In cases where AI tools are used [5], it is crucial to show that the AI did not generate the work independently [5]. The unpredictability of AI outputs undermines claims of human authorship [3], as the USCO asserts that the AI, rather than the artist [3], originated the traditional elements of authorship [3]. The requirement for human authorship applies to all copyrightable works [5], including those made for hire [5], meaning that simply designating AI-generated works as “works made for hire” may not suffice if they lack copyrightability [5]. For authors incorporating AI-generated content [2], when registering works with AI contributions [2], creators must disclose significant AI involvement [2]. This disclosure is particularly important for filmmakers, as it informs distributors and exhibitors about the inclusion of AI-generated material [1], which is not protected by copyright [1]. Copyright protection is granted only for the human contributions and modifications made to AI-generated content [2], and the more an author alters an AI-generated work [2], the greater the likelihood of copyrightability [2]. However, some level of human creativity is essential for a work to be considered copyrightable [2], and using an AI tool developed by the author does not meet this requirement.
To clarify ownership rights associated with AI-generated works [5], parties should include specific assignments in their contracts [5], separate from the works-made-for-hire doctrine [5]. Service agreements often require cooperation from the service provider in registering deliverables with the USCO or other federal agencies [5], but it should be made clear that this cooperation does not guarantee the registrability of the deliverables [5]. Contracts can effectively establish parameters and expectations regarding the use [5], assignment [5], and ownership of AI-generated works [4] [5]. Organizations utilizing AI tools in development should evaluate their contracts to ensure they adequately address these issues [5], while also recognizing that only humans can be acknowledged as authors and hold copyrights under current copyright law. The differing approaches in the US and UK highlight potential policy implications [3], including the risk of overwhelming copyright claims from AI-generated works [3], which could hinder innovation [3]. Some legal theorists suggest placing AI-generated outputs in the public domain to encourage creativity [3], while others argue for recognizing the skilled labor involved in directing AI tools as deserving of copyright protection [3]. The upcoming case Allen v [4]. Perlmutter may further address these ambiguities in the rights of AI creators and human users.
Conclusion
The insistence on human authorship for copyright protection underscores the challenges and complexities introduced by AI in the creative process. While the current legal framework does not recognize AI as an author, it leaves room for human-AI collaborations to be potentially copyrightable, provided there is significant human contribution. This evolving legal landscape necessitates careful consideration of contractual agreements and policy implications to balance innovation with the protection of intellectual property rights.
References
[1] https://www.documentary.org/column/legal-faq-ai-tips-us-documentary-filmmakers
[2] https://www.authorsalliance.org/2023/08/24/copyright-protection-in-ai-generated-works-update-decision-in-thaler-v-perlmutter/
[3] https://pec.ac.uk/blog_entries/copyright-protection-in-ai-generated-works/
[4] https://www.authorsalliance.org/2025/03/19/thaler-v-perlmutter-d-c-court-of-appeals-confirms-that-a-non-human-machine-cannot-be-an-author-under-the-u-s-copyright-act/
[5] https://www.jdsupra.com/legalnews/development-contract-considerations-for-5885869/
