By: Sascha Ghobadi
Volume IX – Issue II – Spring 2024
I. Introduction
In 2024, the rise of generative Artificial Intelligence revolutionized all facets of industry in the United States, yet its immersion into society has resulted in complexities regarding its use in the legal context, more specifically regarding copyright law. Because of AI’s importance in the digital age, it is imperative that the laws surrounding it are without confusion. The incentivization of advancing and integrating AI can only occur if the applicable laws are without confusion. The incentivization of advancing and integrating Artificial Intelligence should be promoted. Therefore, the relevant copyright laws should be clearer in order to promote not only technological achievement but also to ensure an understandable consensus of how AI is protected under copyright law due to the difficulty that arises out of the nuances in the context of Artificial Intelligence. Before discussing the legal issues that arose with the emergence of AI, it is important to understand a brief history of both Artificial Intelligence and United States copyright law.
II. Background on Artificial Intelligence
Artificial Intelligence, or AI, is “technology that enables computers and machines to simulate human intelligence and problem-solving capabilities.” The beginning of the concept of AI can be traced back to Alan Turing’s paper Computer Machinery and Intelligence when he created the “Turing Test”, which constituted a human interrogator attempting to distinguish between a human and computer answer. Turing posed the question: “Can machines think?” [1]
From IBM’s Deep Blue winning a match against world chess champion Gary Kasparov in 1997 to DeepMind’s AlphaGo program beating the world champion Go player in a five-game battle, AI made progress in its capacity for use on an increasingly macro scale.
It was not, however, until 2023 where Artificial Intelligence began to grow rapidly. With the emergence of large language models (LLM), AI entered mainstream discussion. LLMs such as OpenAI’s ChatGPT and Google’s Bard have permitted everyday users to learn from, interact, and create content with Artificial Intelligence. By inputting prompts, users receive a dynamic response generated by the program.
These advancements in AI technology have led to Generative Artificial Intelligence. Generative AI is a form of artificial intelligence that gathers and learns from raw data to create new content ranging from images to music. It makes use of advanced learning machine models to identify patterns in the data. Using machine learning, GenAI can create new instances of data that are similar but not exactly like the original input. [2]
In the United States, copyright law is complex. In the realm of AI, US copyright laws have found issues in their lack of distinctiveness and clarity regarding the legality of ownership when GenAI is involved.
III. Background on Current U.S. Copyright Law
The relevant codes, sections, and laws will be discussed in this section. Code 17, Section 102 of the U.S. Code protects original works of authorship “fixed in any tangible medium of expression.” These works include a broad range of categories from literary, musical, and dramatic pieces to sculptures and architectural designs. Regardless of how it is depicted or described within a copyrighted work, copyright protections do not extend to any idea, process, or operational method. (lexis, a look at future)
Article I, Section 8, Clause 8 of the U.S. Constitution (i.e. Intellectual Property Clause) provides Congress the power “[t]o promote the . . . useful Arts, by securing for limited Times to Authors . . . the exclusive Right to their respective Writings.” [3] Under the Copyright Act, which enacts Congress's authority from the IP Clause, copyright is an intangible right that safeguards original works fixed in a tangible medium of expression. The Copyright Act maintains a criteria for originality that requires the work in question to be independently created with a minimum degree of creativity, a standard that can often be met with ease. To reference an example, the layout of a phone book may qualify, but a simple alphabetical listing does not. Under the Copyright Act, the initial owner of a copyright is the author. The “author” can be defined as the person who created the work and fixed it in the tangible medium. An exception to this section of the law is if a work is a "work-made-for-hire." In these instances, the employer can automatically own the copyright to a work created by an employee within the scope of their employment. This distinction is crucial in determining copyright ownership. The Copyright Act does not explicitly provide definitions of "employee" and "scope of employment", rather these terms are not put forth by the Copyright Act, rather the courts scrutinize these concepts under “the general law of agency,” informed by the Supreme Court in Cmty. for Creative Non-Violence v. Reid. [4] The Copyright Act states that “[c]copyright protection subsists, in accordance with this title, in 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.” [5]
Although the high levels of exposure to AI have allowed for new ways for individuals to create and learn, issues arise in regard to originality and copyright law. The case Thaler v. Perlmutter serves as a representative example of how copyright laws interact with the generative abilities of AI.
IV. Relevant Case: Thaler v. Perlmutter
In Thaler v. Perlmutter, Judge Beryl A. Howell of the U.S. District Court for the District of Columbia ruled on a copyright dispute in regard to a piece of AI-generated artwork titled "A Recent Entrance to Paradise." [6] Stephen Thaler claimed the AI he created called the "Creativity Machine" autonomously created the artwork. Thaler thus sought copyright registration for the work produced. Although the U.S. Copyright Office had already denied Thaler’s application in 2019, citing a lack of human authorship, Thaler challenged this decision as arbitrary under the Administrative Procedure Act. (elaborate on this Act) The court agreed with the Copyright Office and upheld its decision. The court emphasized that copyright law requires some form of human authorship, while also claiming that the law was designed to incentivize human creativity. The judgment was decided using past historical and legal precedents. Through such precedents, the court was able to determine that AI-generated works currently do not meet these criteria. [7]
The key precedents used in the court’s decision were interpretations of the 1909 Copyright Act and the 1976 Copyright Act, which both have consistently held that an "author" in the context of copyright protection must be a human being. The court also referenced the Supreme Court decision in Burrow-Giles Lithographic Co. v. Sarony. These precedents reinforced the argument that copyright is designed to incentivize human creativity, which excludes non-human entities like AI from authorship recognition. This case raises significant questions about the role of AI in creative processes and suggests an evolving legal landscape as AI technologies advance. [8]
V. Issue with Modern U.S. Copyright Law
As evidenced by the Thaler case, current copyright laws need to be expanded and clarified as they pertain to Generative Artificial Intelligence. A significant issue that needs to be considered is how to delineate the degree of human involvement in works generated by AI in order for those works to be afforded protection under the copyright law umbrella. In its final ruling, the Thaler court took care to specify that the scope of its decision was delimited since, although it affirmed that there had to be a de minimis human nexus between the creator and the creation, the facts of the case did not necessitate determining the threshold of human involvement required to qualify a claimant as an author of an AI-generated work. The administrative record the court relied on in reaching its decision indicated that although Thaler had created the AI platform that generated the artwork, Thaler himself admitted to no direct involvement in prompting the production of the artwork by that platform. Nonetheless, the court acknowledged both the evolving nature of AI technology and the commensurate rise of complex legal issues regarding authorship and copyright eligibility. The absence of specificity and clarity in the Thaler case opens the door to exploring ways to meet the challenges of AI-generated works. The Thaler ruling underscores the need for future legislative developments to address the shifting boundaries between human and non-human creative contributions in the context of copyright law. [9]
VI. Final Recommendations
Although it is evident that there are difficulties when applying copyright law to Artificial Intelligence, these laws must be updated or reinterpreted to be more clearly defined. It is important to understand the context regarding why such laws need to be reimagined. This debate over new technologies and copyright law began decades ago in the 1979 CONTU (National Commission on New Technological Uses of Copyrighted Works) report. The CONTU report was commissioned at the behest of Congress to analyze the potential impact new technologies, like computers at the time, had on copyright law. They emphasized that if computer-assisted works met the originality criteria of the 1976 Copyright Act, then such works should deserve protection under the law. However, a key caveat in the report was that computers were tools rather than “authors” with the distinction that there needs to be some level of human creativity involved for the work to qualify for protection. In a 1993 article, renowned Harvard law professor Arthur R. Miller highlighted the CONTU report’s limitations of failing to fully address autonomous AI creations, with his suggesting that copyright law should evolve to include such works if such a technology were to exist. It is clear that emerging technologies may clash with existing copyright laws, as evidenced by the report itself and Miller’s dissent from it. It is also clear that laws are subject to change, especially when they are outdated.
Although not directly about AI, the case Urantia Foundation v. Maaherra established that while human creativity is required for copyright, that criteria could be met through organizing and compiling. In this case, the Ninth Circuit heard a copyright scenario involving a book claimed to be authored by celestial beings. The court declared that although the Copyright Act does not explicitly state that an author must be human, it requires a work’s authorship to have some level of human creativity. The court ruled that the book has the right to be copyrightable as a compilation because there was a level of human effort in the organization and transcription of the celestial messages. The importance of this decision in the context of AI relies upon the reinterpretation of the Copyright Act. As is the case here, there is precedence of updating the law when need be. The decision alludes to the fact that for AI-generated works, human involvement in the programming, creating, and interacting with AI could open the door for a legal argument to be made in its defense of copyrightability as it introduces an observable effort of creativity and organization in the final product. [10]
Cases such as the Thaler case highlight the need for more innovative copyright legislation as it pertains to AI generated content. As it stands, existing copyright law has to play catch up with technological advances, warranting a reevaluation as to how it can be expanded to apply to works whose authorship cannot be credited to human creators in the traditional sense. The Copyright Office, in recognizing the impact the unprecedented pace of rapidly evolving AI platforms will continue to impart on copyright law applications, has committed itself to acknowledging and addressing the burgeoning issues surrounding identifying authorship and conferring copyrights. The US Senate visited this very subject matter in October 2022, acknowledging that the exclusion of AI generated work from the protection of copyright law might deter AI-driven innovations, thus necessitating potential legislative changes in the field. To better accommodate AI-generated works, Senators Thom Tillis and Chris Coons proposed exploring alternative legal frameworks, such as a sui generis classification. Such an approach, where unique legal rights are specifically designed to address circumstances not covered by existing law, could create protections customized for AI innovations, encouraging incentivization while adapting to evolving technological advances. Legal adaptation on this scale would foster a synergistic growth between law and technology, fostering innovation by offering protection of novel AI generated creations. [11]
It is of vital importance that we as a society address how AI interacts with the law. Questions must be asked and answers must be thought about carefully. As the creation of different AI systems becomes more and more prevalent, will the creator of the creator, meaning the AI systems, be able to claim enough of a human nexus to claim authorship of the final creation? The reasoning is that but for the existence of the original creator, the final creation would not have been able to have been generated. As AI continues to not only advance but entrench itself into the fabric of our society, the relevant copyright laws must be reevaluated to allow individuals a clear understanding of the language dictating their rights and protections.
Endnotes
[1] IBM. "What is Artificial Intelligence (AI)?" IBM. 2024. Accessed May 5, 2024. https://www.ibm.com/topics/artificial-intelligence.
[2] Ibid.
[3] ArtI.S8.C8.1 Overview of Congress's Power Over Intellectual Property." Constitution Annotated. United States of America Congress. Accessed May 5, 2024. https://constitution.congress.gov/browse/essay/artI-S8-C8-1/ALDE_00013060/.
[4] Bouffier, Charles, et al. "Generative AIs Challenged By Copyright and Related Rights: A Comparative Approach in European and U.S. Law." LexisNexis. July 1, 2023. Accessed May 5, 2024. https://www.lexisnexis.com/pdf/practical-guidance/ai/generative-ais-challenged-by-copyright-and-related-rights.pdf.
[5] 17 U.S.C. § 102 - U.S. Code Title 17. Copyrights § 102." Codes - FindLaw. January 1, 2024. Accessed May 5, 2024. https://codes.findlaw.com/us/title-17-copyrights/17-usc-sect-102/.
[6] Garbett, Tianna, and James G. Gatto. "Copyright Issues in the World of Gen AI Thaler v. Perlmutter." The National Law Review. October 4, 2023. Accessed May 5, 2024. https://natlawreview.com/article/generative-ai-and-copyright-some-recent-denials-and-unanswered-questions.
[7] "District Court Rules that AI-Generated Works Cannot Be Copyrighted." Morrison Foerster. August 29, 2023. Accessed May 5, 2024. https://www.mofo.com/resources/insights/230829-district-court-rules-that-ai-generated-works.
[8] Ibid.
[9] Grossman, David, and Keane Barger. "Thaler v. Perlmutter." Loeb & Loeb LLP. August 18, 2023. Accessed May 5, 2024. https://www.loeb.com/en/insights/publications/2023/08/thaler-v-perlmutter.
[10] Crouch, Dennis. "AI as Author: Thaler v. Perlmutter Now Before the DC Circuit." Patently-O. April 18, 2024. Accessed May 5, 2024. https://patentlyo.com/patent/2024/04/author-perlmutter-circuit.html.
[11] Bouffier, Charles, et al. "Generative AIs Challenged By Copyright and Related Rights: A Comparative Approach in European and U.S. Law." LexisNexis. July 1, 2023. Accessed May 5, 2024. https://www.lexisnexis.com/pdf/practical-guidance/ai/generative-ais-challenged-by-copyright-and-related-rights.pdf.