On September 26th, 2024, AI artist Jason Allen filed an appeal at the U.S. District Court of Colorado requesting the reversal of a US Copyright Office (‘USCO’) decision to deny copyright protection for his artwork ‘Théâtre d’Opéra Spatial’.
Mr. Allen’s first attempt to secure copyright protection for his work dates back to September 2022, when he filed a registration application with the USCO. Though, in his first application, Mr. Allen did not disclose that his work was created using an AI system (specifically, Midjourney), the USCO was aware of his work and of the involvement of AI since it had garnered national attention in August 2022 for being the first ‘AI-generated’ (or, according to Mr. Allen ‘AI-assisted’) image to win the Colorado State Fair’s annual fine art competition[1].
The decision of the USCO
Following several correspondence exchange with Mr. Allen, the USCO issued its final decision in September 2023. With said decision, the USCO Review Board denied copyright registration for Mr. Allen’s award-winning artwork since it lacks the human authorship necessary to support a copyright claim and, more in depth, ‘it contains more than a de minimis amount of AI-generated content, which must be disclaimed in an application for registration’[2].
In a nutshell, in the USCO’ reasoning[3], if all the traditional elements of authorship, i.e. conception and execution, are generated by a machine, the work lacks human authorship and the USCO cannot register it[4] . If, however, a work containing AI-generated material also contains sufficient human authorship to support a claim to copyright, then the USCO will register the human’s contributions only. In such cases, the applicant must disclose AI-generated content that is more than de minimis. Said disclosure requirement was not met by Mr Allen, who sought to register the entire work and refused to disclaim the portions attributable to AI, hence his work was denied registration.
Mr Allen’s appeal
In the appeal filed last month before the Colorado District Court, Mr. Allen argued that, in finding that his work lacks human authorship, the Office ignored the essential element of human creativity required to create an artwork using Midjourney and ignored the fundamental principle that copyright is meant to be granted to expressions of ideas originating from a human mind[5].
According to Mr. Allen, using Midjourney require the creator to carefully craft prompts to achieve the desired results. In describing the origin of the artwork, Mr. Allen clarifies that the idea behind the artwork was conceived by him, who provided the initial input to the machine, i.e. the image of women in Victorian dresses wearing space helmets. He then selected the colors, the style, the era of the artwork, the setting[6] and arranged all the other elements in the image in order to convey his idea in the artwork.
The final result stems from 624 interactions between Mr. Allen and Midjourney, through which the artist provided new instructions from time to time, based on the AI’s response. Once achieved the desired output, Mr. Allen upscaled and further modified the image for clarity and detail and added few missing elements.
In other words, according to Mr. Allen the work was not created by Midjourney merely through inputting a few prompts or pressing a button. The prompt serves as the initial creative input, embodying the human’s original idea and vision for the artwork and the AI system only functions as a tool to assist the human creator in the actualization of his/her creative idea[7].
The underlying AI-generated work merely constitutes raw material, which Mr. Allen has transformed through his artistic contributions. Said contributions are on par with that expressed by other types of artists in artworks deemed capable of copyright protection by the USCO. Therefore, regardless of whether the underlying AI-generated work is eligible for copyright registration, the entire work in the form submitted to the copyright office should be accepted for registration.
In this regard, Mr. Allen argues that the statutory language of the U.S. Copyright Act supports the notion that AI-assisted artworks with substantial human creative input can indeed be eligible for copyright protection[8] and that the requirements of human conception and execution requested by the USCO are met in his artwork, even if they are declined differently than in traditionally human-created artworks before generative AI came into play.
Finally, Mr. Allen stated that, by refusing to register contents generated via Midjourney and other generative AI systems, the USCO has failed to remain technologically neutral and that denial of copyright protection for the output of such tools would result in a void of ownership: ‘In many instances, a Copyright Examiner may not even be able to distinguish an artwork that used AI tools to assist in the creation from one which does not use any computerized tools, thus making the review process entirely arbitrary’[9].
The USCO denial, according to Mr. Allen, is also causing him substantial damages, since, in the months following the final USCO decision, consistent infringements of his work has jeopardize Mr. Allen’s ability to monetize his artistic creation[10].
Generative AI and copyright protection (brief considerations)
The approach adopted by the USCO in Mr. Allen’s case is not novel and, on closer inspection, it is highly common in cases involving applications for copyright registration of artworks generated employing AI[11]. However, this approach, whereby AI-generated works should not be granted protection under copyright law as it was originally conceived to protect human creativity, seems tautological: at the time when copyright protection arose, human creativity was the only conceivable form of creativity, whereas the idea that a machine could create artworks independently (or with limited human contribution) was unimaginable.
Also, the approach whereby works involving AI should be granted protection or not depending on the level of human intervention detectable in them (shown by the USCO in the case at hand) is indeed consistent with traditional concepts of copyright law. However, the same is hardly applicable in practice, as it is based on uncertain parameters (likewise the minimum human contribution to support a copyright claim) that harbour excessive discretion, as argued by Mr. Allen.
This approach is also not consistent with the current technical landscape, whereby AI systems become increasingly autonomous and sophisticated every day, requiring, as a result, less and less human involvement. Nevertheless, applications for registration of AI-generated works are steadily increasing at copyright offices worldwide, calling for a re-thinking of some traditional copyright concepts, so to adapt them to technological evolution[12].
Finally, the approach outlined above, although with the stated purpose of safeguarding human creativity as the bedrock of copyright, ultimately implies the fall of the work into the public domain thus preventing the human artist from exploiting the artwork (as Mr. Allen’s case), whereas works created without employing AI, yet with a degree of creativity equal to pressing a button, would be unquestionably protectable under current USCO case law[13].
The granting of IP protection to AI-generated works does not imply, per se, the awarding of any rights to a machine. Based on the economic rationale of copyright, the incentive provided by IP rights shall be granted to those who are in a position to invest in artificial innovation (thus, clearly, human beings), so as not to hinder, rather to encourage overall progress and follow the path indicated by the European Directives on copyright, most recently the ‘Copyright Directive’[14].
The decision of the Colorado District Court in the case of Mr. Allen is forthcoming. The same may either follow the USCO’s reasoning and approach, as done by the Columbia District Court in Mr. Thaler’s case[15], or favouring a more progressive and realistic approach, as other courts have done in some recent landmark rulings[16].
[1] Contrary to his registration application to the USCO, in his application to the Colorado State Fair’s competition Mr. Allen disclosed that the work had been created using Midjourney (since he was competing in the digital art category). However, none of the jurors was aware that Midjourney is a generative AI system.
[2] See U.S. Copyright Office Review Board, ‘Second Request for Reconsideration for Refusal to Register Théâtre D’opéra Spatial (SR # 1-11743923581; Correspondence ID: 1-5T5320R)’ of 5th September 2023, paragraph III, pg. 3, available at the following link: chrome-extension://efaidnbmnnnibpcajpcglclefindmkaj/https://www.copyright.gov/rulings-filings/review-board/docs/Theatre-Dopera-Spatial.pdf .
[3] Following the reasoning already provided in a public guidance on registration of works created by a generative AI systems issued by the USCO in March 2023 (see Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence, 88 Fed. Reg. 16,190, 16,192, 16th March 2023, at the following link: https://www.federalregister.gov/documents/2023/03/16/2023-05321/copyright-registration-guidance-works-containing-material-generated-by-artificial-intelligence ).
[4] This conclusion is drawn by the USCO based on the fact that the U.S. Copyright Act protects, and the USCO registers, ‘original works of authorship fixed in any tangible medium of expression’ (see 17 U.S.C. § 102a). Courts have interpreted the statutory phrase ‘works of authorship’ to require human creation of the work, stating that ‘human authorship is a bedrock requirement of copyright’ (see the decision issued by the U.S. District Court of Columbia in Thaler v. Perlmutter case, Civil Action No. 22-1564 on 18th August 2023, pg. 8). For this reason, courts have uniformly rejected attempts to protect the creations of non-humans through copyright (see U.S. Copyright Office Review Board, ibidem, paragraph III(A), pg. 3).
[5] See Plaintiff’s complaint and request for declaratory relief and demand for jury trial (‘Plaintiff’s complaint’), Civil Action No. 1:24-cv-2665, filed on 26th September 2024 at the U.S. District Court of Colorado by Tamara Pester, Esq. on behalf of Mr. Allen in the Allen v. Perlmutter case, paragraph 69, available at the following link: chrome-extension://efaidnbmnnnibpcajpcglclefindmkaj/https://fingfx.thomsonreuters.com/gfx/legaldocs/gdvzkrrmapw/AI%20COPYRIGHT%20REGISTRATION%20appeal.pdf .
[6] I.e. a ‘grand theater, with an audience watching intently, overlooking a large circular window through which the vast expanse of the outer world is visible, adding an otherworldly ambiance to the performance’ (see Plaintiff’s complaint paragraph 16).
[7] In this regard, in the Plaintiff’s complaint Mr. Allen compares the creative process employing generative AI to photography: ‘In a literal sense, the camera captures and develops the photograph, yet no one would dispute the photographer’s authorship. Similarly, Mr. Allen conceived of the image, developed the prompt, and made adjustments to refine the output, thereby ensuring the final image reflects his mental vision. The AI system, much like a camera, is a tool through which the artist’s creativity is realized in visual form.’ (see Plaintiff’s complaint, ibidem, paragraph 80). When photography first gained popularity, critics argued that it lacked skill and artistry, yet it has since become a highly respected and valued art form. According to Mr. Allen AI-assisted art holds the potential to do the same and this evolution should be embraced as a positive development in the creative landscape (see Plaintiff’s complaint, ibidem, paragraph 73).
[8] See 17 U.S.C. § 102(a).
[9] See Plaintiff’s complaint, ibidem, paragraph 16.
[10] The following link provides an example of a canvas featuring Mr. Allen’s artwork available for sale on Etsy for $18.65: https://www.etsy.com/listing/1358054636/theatre-dopera-spatial-canvas-art-print .
[11] E.g. in the case of Mr. Thaler, who in February 2022 had his application for copyright registration of the work ‘A Recent Entrance to Paradise’ rejected (see USCO, Second Request for Reconsideration for Refusal to Register A Recent Entrance to Paradise (Correspondence ID 1-3ZPC6C3; SR # 1-7100387071) of 14th February 2022). Unlike Mr. Allen, Mr. Thaler claimed that the AI itself conceived of and created the work independently, with him seeking copyright protection on that AI-generated work based on various legal theories under which a copyright in the AI’s work would transfer to him, e.g., as inventor of the AI-author (called ‘Creativity Machine’) or as the hardware’s owner.
[12] In line with the ‘Constitutional mandate to promote the of science and the useful arts (see Plaintiff’s complaint paragraph 96).
[13] ‘Courts have found it is now settled beyond question that practically anything novel can be copyrighted, even if there is only a faint trace of originality […]. All that is needed to satisfy both the Constitution and the statute is that the ‘author’ contributed something more than a ‘merely trivial’ variation, something recognizably his own. […] No matter how poor artistically the ‘author’s addition, it is enough if it be his own’ (see Plaintiff’s complaint paragraph 46 and the case law mentioned thereby).
[14] Directive (EU) 2019/790 of 17 April 2019.
[15] See U.S. District Court of Columbia, ibidem, where the Court upheld the USCO’s decision and denied copyright protection to Mr. Thaler’s artwork, as described above.
[16] In November 2019, the Court of Nanshan (China) granted copyright protection to a newspaper article written by a robot called ‘Dreamwriter’, owned by the company Shenzhen Tencent, and ordered the other party, Shanghai Yingxun, to pay damages for copyright infringement. In this case, the Nanshan Court found that the AI-written article had a thoughtful structure, clear logic and a certain degree of originality, which met all the requirements for copyright protection. In this work, the human hand was identified as that of the team of Shenzhen Tencent engineers who: designed the operation of the algorithm; selected and entered the data into the algorithm; set the goal (see Shenzhen Nanshan District People’s Court, Shenzhen Tencent Computer System Co., Ltd. vs. Shanghai Yingmou Technology Co., Ltd., available at this link: http://www.ecns.cn/news/2020-01-09/detail-ifzsqcrm6562963.shtml).