Introduction
The legal landscape surrounding copyright protection for works generated by artificial intelligence (AI) is evolving, with recent court rulings reinforcing the necessity of human authorship for copyright eligibility. This ongoing debate highlights the challenges and implications of AI-generated content within existing intellectual property frameworks.
Description
On March 18, 2025 [3] [5] [6], the US Court of Appeals for the District of Columbia Circuit unanimously upheld previous rulings that original artwork created solely by artificial intelligence (AI) without any human contribution is not eligible for copyright registration under the Copyright Act of 1976. This decision reinforces the established principle that copyright is reserved for works produced by human authors [6], as the law mandates human authorship for copyright protection [3]. The case centered on Dr. Stephen Thaler and his AI system, DABUS [3] [4] [5] [7], which generated a piece titled “A Recent Entrance to Paradise.” Thaler’s application for copyright registration, initially submitted in 2018, was rejected by the US Copyright Office in 2022 [6], which emphasized the necessity of human involvement in the creation of works. The US Copyright Office has also denied copyright claims for images generated by another AI system [1], Midjourney [1], affirming that works created by AI without human input cannot be copyrighted under US law [1].
The court [1] [3] [4] [5] [6] [7], led by Judge Patricia A [3]. Millett [3], affirmed that the language of the Copyright Act presupposes human authorship, defining an “author” as a human being [3]. The ruling highlighted that machines lack the attributes associated with authorship [3], such as property rights and human lifespans [3]. Thaler’s attempt to frame the AI as an “employee” was also rejected [5], with the court clarifying that an AI cannot enter into contracts [5], thus precluding any employment relationship [5]. While the court did not address whether Thaler could claim authorship based on his role in creating DABUS [3], it acknowledged that the registrability of AI-generated works depends on the specific circumstances of how the AI was utilized [3], including the extent and quality of human contribution [2]. Minor adjustments to AI output may not suffice for copyright eligibility [2], whereas substantial creative involvement can lead to protection for the human contributions [2].
This decision aligns with US patent doctrine [3], which similarly requires human inventorship for patent protection [3], as illustrated in Thaler v. Vidal [3]. The rationale for requiring human authorship and inventorship stems from the belief that only humans can truly “create” art or conceive inventions [3]. A White Paper from the University of Oxford supports this notion, outlining that creativity encompasses external [3], mental [3], and social components [3], with generative AI currently lacking the mental component essential to human creativity [3].
Moreover, the White Paper suggests that generative AI may adversely affect human creativity by promoting a focus on speed and passive consumption of AI-generated outputs [3], potentially leading individuals to doubt their creative abilities [3]. The ongoing debate surrounding Dr [3]. Thaler and DABUS challenges existing intellectual property doctrines [3], raising questions about the copyrightability of computer-generated works in the modern context [3]. While some advocate for the protection of AI outputs based on their quality [3], courts maintain that human involvement is typically present in AI-generated works [3], whether through the creation of the AI [3], goal-setting [3], or evaluation of outputs [3].
In Australia [6], copyright laws similarly necessitate human authorship for a work to qualify for protection [6], as outlined in the Copyright Act 1968 (Cth) [6]. Consequently [6], works created entirely by AI without direct human involvement may not be eligible for copyright under current legislation [6]. The discourse surrounding AI and copyright in Australia is evolving [6], with intensified debates among legal experts [6], artists [6], and industry stakeholders regarding the implications of AI-generated works [6]. Key issues include the application of current laws to AI training and operation [6], particularly concerning ownership and intellectual property rights [6]. A major concern is the use of copyrighted material in AI training datasets [6], with some artists arguing that AI programs exploit their work without permission [6], sparking discussions about potential copyright infringement and the ethical considerations of AI-generated content [6].
As AI technology continues to advance, the legal framework governing its role in creative industries is expected to be a focal point for ongoing discussions [6], prompting policymakers to consider legislative updates to balance innovation with creator protections [6]. The distinction between human creativity and AI capabilities may blur [3], presenting new challenges in intellectual property law and prompting further legal cases that will test the boundaries of existing laws and potentially reshape the framework governing intellectual property in the digital age.
Conclusion
The court’s decision underscores the importance of human authorship in copyright law, reflecting broader concerns about the implications of AI in creative fields. As AI technology evolves [2] [3] [5] [6], it will be crucial for legal frameworks to adapt, ensuring a balance between fostering innovation and protecting the rights of human creators. This ongoing dialogue will likely shape the future of intellectual property law, as stakeholders navigate the complexities introduced by AI-generated content.
References
[1] https://www.straitstimes.com/world/united-states/us-appeals-court-rejects-copyrights-for-ai-generated-art-lacking-human-creator
[2] https://www.lawyer-monthly.com/2025/03/court-rules-ai-cannot-be-copyrighted-landmark-ruling-on-human-authorship/
[3] https://natlawreview.com/article/dc-circuit-denies-copyright-ai-artwork-what-humans-have-and-artificial-intelligence
[4] https://techvibe.ai/us-court-says-ai-generated-art-cannot-have-copyright/
[5] https://kotopat.com/en/can-ai-hold-copyright-the-intersection-of-law-and-creativity-in-a-u-s-federal-court-ruling/
[6] https://www.capturemag.com.au/news/us-appeals-court-rules-ai-generated-works-ineligible-for-copyright-protection
[7] https://www.jdsupra.com/legalnews/d-c-circuit-denies-copyright-to-ai-4164817/
