USCO Denies Copyright Registration for AI-Generated Work

In the ever-evolving landscape of creative digital works, the intersection of artificial intelligence and copyright law is becoming a pivotal point of contention. The U.S. Copyright Office (USCO) has once again entered the spotlight with their steadfast stance on AI-generated works. On December 11, they released a public letter reaffirming their position by refusing copyright registration for an AI-generated piece—the fourth decision of its kind—sending ripples across the AI news industry.

The work in question is a digital image titled “SURYAST,” a blend of a sunset photograph and the iconic style of Vincent van Gogh’s “The Starry Night,” crafted in part by a custom-developed software named RAGHAV. Animator Ankit Sahni, who submitted an application for copyright in December 2021, cited both himself and the advanced AI tool as co-authors. This sparked an intensive review process, during which the USCO dug deep into the creative process behind the image.

Sahni’s initial application raised eyebrows at the USCO. In June 2022, it received an initial refusal due to the inability to discern human authorship within the piece. The USCO underscored the challenge of delineating between the human contribution and the computer-fabricated final product, a necessity for a work to be considered a derivative piece authored by a human.

Persistent in his claim, Sahni applied for reconsideration in September 2022, challenging the notion that human authorship must equate to full, direct creation without any assistance. However, the USCO wasn’t swayed, upholding their refusal. Another plea for reconsideration came in July 2023, this time Sahni contended that RAGHAV was simply an “assistive software tool,” likening the AI’s role to the initial sketching of a painter or the forming of clay by a sculptor—preliminary steps rather than complete authorship.

Sahni’s defense primarily revolved around three points: the purported minimal involvement of RAGHAV, the human-authored aspects of SURYAST, and the argument that the AI-generated work wasn’t derivative due to its lack of substantial similarity to the initial photograph and van Gogh’s art.

Despite these assertions, the USCO was not convinced, noting inconsistencies between Sahni’s provided descriptions and the apparent functionality of RAGHAV. They emphasized that the way RAGHAV reinterpreted the photograph in the light of another painting’s style was intrinsic to RAGHAV’s programming and training images, not specific instructions or contributions by Sahni.

A significant portion of the USCO’s decision rests upon the de minimis standard, a principle derived from the U.S. Supreme Court’s Feist v. Rural Telephone decision. According to this standard, which forms part of the USCO’s Copyright Registration Guidance for AI-generated works, any AI-generated content that is more than de minimis must be disclosed and excluded from copyright applications. Works predominantly generated by AI fall into the public domain.

This is not an isolated case. An earlier documented refusal to recognize AI authorship concerned “A Recent Entrance to Paradise,” a piece by Dr. Stephen Thaler, who sought registration with the AI as the author. This refusal was challenged unsuccessfully in court.

The partial cancellation of a registration for a graphic novel by Kris Kashtanova also gained attention. The novel utilized Midjourney, an AI images generator tool, which prompted the USCO to call into question registrability due to the use of “randomly generated noise.”

Moreover, the USCO has withheld copyright registration for Théâtre D’opéra Spatial, an AI video generator work that garnered attention after its win at the 2022 Colorado State Fair’s digital arts category. It’s noteworthy that Sahni’s attempts aren’t limited to the U.S. In 2020, the SURYAST was granted copyright in India.

These decisions highlight the complex relationship between copyright laws and the burgeoning field of AI-created art. As creators increasingly utilize AI tools, whether for generating AI images or analyzing writing with an AI text generator, the legal community and the public alike are being propelled into ongoing discussions and debates. The intersection of artificial intelligence, creative expression, and copyright legislation will surely continue to be a hot topic within the latest AI news & AI tools sector. Industry watchers and creative professionals must stay informed, as regulatory perspectives and the capabilities of generative AI evolve.