AI-GENERATED WORKS: WHO OWNS THE COPYRIGHT?
Our previous articles laid the foundation for understanding the profound connection between AI and creativity in the ever-evolving world of AI-generated art. But the story doesn’t stop there; it’s a journey that stirs curiosity and concern. In a focused exploration of US and Chinese law, we are set to uncover the captivating intricacies of AI-driven copyright — an issue that holds profound significance for artists, writers, marketers, and businesses alike. Join us in unravelling the vital authorship and ownership aspects in this ever-evolving landscape.
INTRODUCTION
In an era where artificial intelligence (AI) has become a prolific creator of visual content, the question of copyright ownership in AI-generated works has emerged as a pivotal legal and ethical dilemma. This phenomenon marks a significant departure from traditional paradigms where human creativity was the sole progenitor of artistic works.
The heart of this issue lies in the interpretation of copyright laws, initially designed in a pre-digital age, primarily focused on human authorship. The advent of AI challenges these foundations, bringing forth scenarios where the ‘creator’ is a non-human being.
This article delves into the complexities of copyright law as applied to AI creations, comparing and contrasting approaches in different jurisdictions, focusing on the United States and China. These two countries offer contrasting perspectives and legal frameworks, providing a unique lens through which to examine AI’s broader implications in intellectual property rights.
UNITED STATES
The United States’ position on copyright in the context of AI-generated works is unequivocally rooted in the principle that copyright ownership is exclusively reserved for humans. Let’s explore why.
The US Copyright Act of 1976 specifies that copyright protection is immediately granted upon the creation of “original works of authorship fixed in any tangible medium of expression, now known or later developed, from which they can be perceived, reproduced, or otherwise communicated, either directly or with the aid of a machine or device” (17 U.S.C. § 102(a)).
Crucially, the Copyright Act mandates that the “fixing” of the work in a tangible medium must be done “by or under the authority of the author” (17 U.S.C. § 101). Therefore, to qualify for copyright protection, a work must have an identifiable “author.”
In the landmark 1884 case of Burrow-Giles Lithographic Co. v. Sarony (111 U.S. 53, 58), the US Supreme Court addressed the question of whether photographs could be considered copyrightable works of “authors,” despite being produced by a mechanical device like a camera. The Court concluded that photographs do indeed qualify as copyrightable creations because they embody the “original intellectual conceptions of the author” (Sarony, 111 U.S. at 59).
This significant U.S. copyright law case is based on the following story. Napoleon Sarony, a prominent photographer of his time, had taken a photograph of the famous author and playwright Oscar Wilde. Burrow-Giles Lithographic Co. reproduced Sarony’s photograph without his permission and argued that photographs were not eligible for copyright protection because they were not considered “writings” within the meaning of the Copyright Act 1870.
The Court reasoned that while a camera may only provide a “mechanical reproduction” of a scene, the photograph is ultimately the result of a photographer’s “mental conception.” This conception is realised through various creative decisions: the posing of subjects, the selection and arrangement of costumes and accessories, the crafting of graceful outlines, and the manipulation of light and shade. The photographer’s choices in evoking expression and the overall arrangement and representation of the subject are vital factors that transform a mere mechanical reproduction into an original work of art.
This ruling underscored the importance of human involvement and creative control in determining whether a new type of work falls within the bounds of copyright. The Court recognised that even when a work is produced with the aid of a mechanical device, it is the human element — the photographer’s creative vision and choices — that imbue the work with the originality necessary for copyright protection.
In the well-known case of Naruto v. Slater (888 F.3d 418, 420, 9th Cir. 2018), the Ninth Circuit Court addressed an unusual scenario involving a crested macaque who had taken photographs of itself. The court ruled that the monkey, a non-human animal, could not sue under the Copyright Act, stating that “all animals, since they are not human, lacked statutory standing under the Copyright Act”.
While the decision primarily focused on the issue of standing rather than directly addressing the copyrightability of the photographs taken by the monkey, the Naruto court indirectly touched upon the fundamental question of whom the Copyright Act is intended to protect. In this context, the court concluded that the Act protects human authors, implying that authorship and the right to seek protection under copyright law are exclusively human privileges.
Finally, in the very recent United States District Court for the District of Columbia Thaler v. Perlmutter et al, №1:2022cv01564 — Document 24 (D.D.C. 2023) case, Judge Beryl A. Howell stated:
“Copyright is designed to adapt with the times. Underlying that adaptability, however, has been a consistent understanding that human creativity is the [condicio] sine qua non at the core of copyrightability, even as that human creativity is channelled through new tools or into new media” — Beryl A. Howell, United States District Judge, Memorandum opinion, Aug 18, 2023
This emphasis on human creativity as the essential condition of copyrightability becomes particularly relevant in the context of AI-generated content, where the definition and role of the “author” are subjects of legal and philosophical debate.
The essence of the discussion is that similar to the Sarony case, where human creativity was key to copyrighting a photograph created with a mechanical device, the role of human input is crucial in AI-generated images. The analogy suggests considering AI as a modern mechanical device, where the pivotal question is the extent of human involvement in the creative process. The level of human influence in AI-created works determines if the AI system’s user can be recognised as the “author” of the resultant work.
This evolving landscape raises complex questions: How much human creativity is essential for a work to be copyrightable when generated by AI? What is the scope of protection for images produced by AI, primarily when the AI’s training may involve pre-existing works whose origins are unknown? Additionally, how should copyright law adapt to encourage and regulate the use of AI in creative processes? These considerations highlight the need to reassess traditional notions of authorship and originality in the face of advanced technologies that increasingly distance human creators from the direct creation of the final work. This reevaluation is crucial in defining the future of copyright in an era where AI plays a significant role in artistic and creative endeavours.
In its decision in 2022, the US Copyright Office addressed the intricacies of copyright in the realm of AI-generated materials. The case involved a graphic novel that blended human creativity with AI technology.
Initially, Ms Kashtanova received a copyright registration for her graphic novel, “Zarya of the Dawn,” which incorporated both her creative work and images generated by the MidJourney AI system.
Upon reevaluation, the Copyright Office focused on the extent of Kashtanova’s involvement in creating the AI-generated images. They scrutinised how the images were produced — Kashtanova provided descriptive prompts to Midjourney, who then autonomously generated the images. This process, the Office determined, did not reflect sufficient human creativity or control from Kashtanova’s side, particularly regarding the “traditional elements of authorship.”
As a result, the Office concluded that Midjourney, rather than Kashtanova, was the driving force behind the creation of these images. The original method used by Kashtanova was considered unpredictable and not indicative of her being the “mastermind” behind the images, a key criterion for copyright eligibility.
Consequently, the original copyright registration for the graphic novel was revised. The updated certificate recognised only the components directly authored by Kashtanova, primarily the novel’s text. This decision underscored a critical point in copyright law concerning AI: while AI-generated elements in a work are not eligible for copyright protection, human-authored parts, such as Kashtanova’s text, remain protectable. This clarification by the Copyright Office delineates the boundaries of copyright in the evolving landscape of AI-assisted creative works.
CHINA
The Beijing Internet Court made an unexpected ruling on November 27, 2023, regarding the copyright infringement of AI-generated images.
In this case, Mr. Li (plaintiff) used the AI software Stable Diffusion to create an image and publish it online. The image was later used without permission by Ms. Liu (defendant). Upon discovering the infringement, Mr. Li initiated legal proceedings against Ms. Liu, alleging copyright infringement and seeking both a formal apology and monetary compensation of 5,000 yuan. In her defence, Ms. Liu contended that she needed to be more specific about Mr. Li’s ownership rights over the disputed image.
The court ruled in favour of the plaintiff, recognising the AI-generated image as a copyrighted work. However, the Beijing Internet Court eventually ordered Ms Liu to pay Mr Li 500 yuan (Approximately 500 Yuan is equivalent to 75 USD and 60 GBP) in compensation, alongside issuing a public apology.
“… the Court held that:
1. The image involved in the case meets the requirements of “intellectual achievements”.
2. It meets the requirements of “originality”.
3. It should be recognised as a work and protected by copyright laws.
4. It is a work of art.
5. The plaintiff is the author of the image and enjoys the copyright thereof”. — Aris Teon, AI-Generated Images: Who Owns the Copyright? Landmark Ruling From a Chinese Court, Nov 23, 2023
Being involved in designing the presentation of characters, selecting prompt words, arranging the order of these words, and setting relevant parameters, the plaintiff “made a certain amount of intellectual investment” needed for the image to meet the legal requirements of “originality”. Therefore, the court found that these activities constituted a significant intellectual input by the plaintiff, qualifying the images as original works of authorship.
CONCLUSION
The contrast between the US and China’s approaches regarding the role of AI in creative works is striking. The US Copyright Office maintains a traditional view with a strict definition of authorship, insisting on direct human creation and intellectual input for recognition under copyright law. Conversely, China adopts a more flexible approach. It recognises AI as a legitimate tool in the creative process, implying a broader interpretation of authorship.
While it may seem that China welcomes innovation, the US approach appears to place a higher value on human personality within intellectual property (IP) by imposing stricter rules on what constitutes copyrightable material. This perspective emphasises that significant human effort and creativity are prerequisites for a work to be granted IP protection. This stringent criterion reflects a deep appreciation for individual creativity and the unique contributions of the human mind, suggesting that the US legal system prioritises the recognition and protection of personal creative labour and intellectual investment in the IP landscape.
This article was written by Liza Lobuteva & Yuriy Brisov of Digital & Analogue Partners. Visit dna.partners to learn more about our team and the services.
Be digital, be analogue, be with us!
