A Rose Enigma: Pending (Right) Protection

December 8, 2023

By Atreya Mathur

Introduction: From Zarya of the Dawn to Rose Enigma

The US Copyright Office (USCO) issued a pivotal decision regarding the copyright registration of “Zarya of the Dawn,” a graphic novel that incorporates artificial intelligence (AI) in the image creation process. Authored by Kris Kashtanova, the graphic novel utilized the Midjourney AI platform to generate some of its illustrations, sparking a debate about the intersection of AI-assisted art and creativity and copyright protection.

Kashtanova successfully obtained a copyright registration for the entire graphic novel in September 2022. However, a subsequent review by the USCO, initiated in October 2022, led to a reevaluation of the registration. The USCO’s final decision, issued on February 21, 2023,^[1] resulted in a limited registration that covered the text, selection, and arrangement of the images, but notably excluded copyright protection for the individual AI-generated images themselves.^[2]

The Issue with Zarya of the Dawn

The crux of the matter lies in the USCO’s stance that copyright protection should be exclusive to human authors. The decision hinged on the perceived lack of substantial creative control exerted by Kashtanova over Midjourney’s output, emphasizing the “distance” between user input and AI-generated output.^[3] Drawing parallels to commissioning a visual artist, the USCO distinguished Midjourney from traditional tools like cameras or software, asserting that the AI platform produces less predictable and controllable results.^[4]

Regarding the images independently, the Office concluded that images generated by Midjourney lacked originality and protection.^[5] The USCO relied on the principle that works produced by a machine without human creative input are ineligible for copyright. While Kashtanova claimed to guide the creative process, it was clarified that Midjourney, not Kashtanova, originated the “traditional elements of authorship” in the images.^[6] Kashtanova’s involvement was described as selecting, tweaking prompts, and engaging in trial-and-error, but she does not claim to create visual material herself. The final images are to be attributed to Midjourney’s autonomous generation.^[7] So while the selection and arrangement is protected, the images in itself are not protected by copyright.

Legal experts, artists, and AI developers have been polarized in their reactions to the decision. Some see it as a partial victory, acknowledging Kashtanova and Midjourney’s role in the creative process while retaining certain rights. Others, however, view it as a setback for AI-assisted creativity, arguing that it neglects protection for potentially original and expressive AI-generated images. The decision’s perceived narrowness has ignited discussions on its adaptability to the diverse landscape of AI platforms and their intricate interactions with human users.

The Decision

While recognizing Kris Kashtanova’s authorship of the text and the overall structure of “Zarya of the Dawn,” the USCO’s decision underscores the non-human authorship of the AI-generated images. The decision prompts broader questions about the future legal landscape of AI-generated works. As AI continues to permeate various creative domains, challenges arise regarding music, poetry, code, and other forms of expression. Determining the level and nature of human involvement and authorship in these works, and striking a balance between the interests of human creators, AI developers, and the public, presents a challenge. Moreover, the decision raises concerns about how legal frameworks will keep pace with the evolution and innovation of AI technology.

What’s the Enigma?

Following the USCO’s partial cancellation of Kashtanova’s registration for Zarya of the Dawn, the artist remained undeterred and submitted a new application for a distinct artwork titled “Rose Enigma.” This piece, produced using AI tool Stable Diffusion, incorporates Kashtanova’s hand-drawn art and other input, emphasizing the artist’s creative control. Kashtanova’s lawyers emphasized Kashtanova’s contributions, including a pen drawing, specific text prompts, and the use of a separate tool to control the viewer’s perspective, as evidence of the artist’s substantial creative input.^[8]

Kashtanova’s attorneys argued that “Rose Enigma” was not a random creation but a result of intentional choices and creative control, positioning Kashtanova as the author. The application asserts that the artist’s exercises of control and creative decisions brought visible form to the work, aligning with copyright eligibility criteria.^[9]

Kashtanova selected a cyborg as the subject for “Rose Enigma” and took active steps to guide Stable Diffusion in realizing this choice.^[10] Through textual prompts, Kashtanova included specific details like “cyborg,” ensuring the accurate rendition of their envisioned subject. Kashtanova had a distinct conception of how the cyborg should appear—a young, human-looking woman with flowers emerging from her head. To convey these details, Kashtanova added descriptors such as “young,” “woman,” and “flowers coming out of her head” to the textual prompt. Not content with generic flowers, the artist also specified roses for the depiction. Using parentheses around “roses” in the prompt, with a multiplier effect, Kashtanova ensured Stable Diffusion focused on this particular floral element, adding nuance to the artistic expression. To capture the desired style, a blend of photorealism and hyperrealism, Kashtanova instructed Stable Diffusion with terms like “photorealism,” “hyper realism,” “8k,” and “hyper-detailed” in the prompt. This conveyed Kashtanova’s intention for the work to resemble a highly realistic painting while incorporating the emotive quality found in hyperrealist pieces. Kashtanova’s vision extended to dramatic lighting, characterized by harsh shadows—a preference rooted in their experience as a professional photographer. By including “cinematic lighting” in the textual prompt, Kashtanova aimed to exert control over Stable Diffusion, ensuring the realization of their envisioned lighting effects.^[11]

Heather Whitney, an attorney for Kashtanova, stated that they believe “Rose Enigma” is eligible for registration under the new guidance provided by the Copyright Office. The guidance emphasizes the disclosure of AI use while maintaining the copyright eligibility of the artist’s contributions.

In addition, the artist states that the “goal” of “Rose Enigma” is to provide education to the US Copyright Office on the diverse applications of A.I. technology.

Conclusion

Considering the detailed artistic direction provided by Kashtanova for the creation of “Rose Enigma” using Stable Diffusion, the potential for copyright registration becomes even more nuanced. While Kashtanova actively guided the AI through specific prompts, shaping the subject, floral elements, style, and lighting, the question of copyright eligibility hinges on the extent of human input and creative control. Kashtanova’s meticulous choices, from the cyborg’s appearance to the nuanced details of roses and stylistic preferences, demonstrate a deliberate and thoughtful artistic direction. However, the AI’s role in generating the images and the extent to which it “autonomously contributes” to the creative process raise complexities in the copyrightability analysis.

In the context of copyright law, the ability to claim authorship and secure copyright protection typically rests on human creativity and originality. While Kashtanova’s influence is evident in the detailed prompts and artistic preferences, the underlying generation process by Stable Diffusion introduces an element of “machine autonomy.”^[12]

The Copyright Office’s approach to AI-generated works involves assessing the level of human input and intervention. Kashtanova’s active role in directing Stable Diffusion provides a compelling argument for copyright eligibility, especially considering the specificity of the prompts and the unique artistic vision conveyed. However, the inherent unpredictability of AI-generated outcomes and the absence of precise control over each detail may present challenges to a straightforward copyright claim.

Kashtanova’s persistence in seeking copyright protection for AI-generated art raises questions about the evolving landscape of copyright law in the realm of AI-assisted creativity. As legal discussions continue, artists like Kashtanova navigate the balance between AI assistance and maintaining authorship rights over their creative expressions.

About the Author

Atreya Mathur is the Director of Legal Research at the Center for Art Law. She was the inaugural Judith Bresler Fellow at the Center (2021-22) and earned her Master of Laws from New York University’s School of Law where she specialized in Competition, Innovation, and Information Laws, with a focus on copyright, intellectual property, and art law.

Sources

1. United States Copyright Office, Registration # VAu001480196, available at https://www.copyright.gov/docs/zarya-of-the-dawn.pdf ↑
2. Id. ↑
3. Id. ↑
4. Id. ↑
5. Id. ↑
6. Id. ↑
7. Id. ↑
8. Registration Cover Letter for Rose Enigma, available at https://drive.google.com/drive/u/0/folders/1MOZe5wLRJF3hR6I-x77GXeuTG-QAUv37 ↑
9. Also see Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence, 88 Fed. Reg. 16,190 (Mar. 16, 2023) (to be codified at 37 C.F.R. § 202) ↑
10. Registration Cover Letter for Rose Enigma, available at https://drive.google.com/drive/u/0/folders/1MOZe5wLRJF3hR6I-x77GXeuTG-QAUv37 ↑
11. Id. ↑
12. Id. ↑