Artificial Intelligence and the Intellectual Property Rights Debate

What does ChatGPT have in common with Naruto the macaque? At first glance, a sophisticated artificial intelligence (AI) language model and a monkey might seem worlds apart. However, both have been central figures in discussions about the nature of authorship and copyright. Naruto became famous for taking a selfie with a photographer’s camera, leading to a legal debate about whether a non-human can hold a copyright. Similarly, as systems like ChatGPT produce content, questions about the eligibility of machine-generated outputs for copyright protection are popping up left and right. In both cases, the underlying issue is the evolving definition of authorship in an age where traditional boundaries of creativity are being challenged.

In today’s ever-evolving digital landscape, generative AI is not just a technological wonder but a critical business consideration. It’s no longer just about coding and algorithms; it’s about reshaping entire industries, navigating new legal terrains, and redefining what we perceive as ‘creation’. Businesses are adopting and incorporating generative tools into their activities at a break-neck pace, but the legal community is still grappling with major questions about the reconciliation of AI and the preexisting legal frameworks that determine how a business can protect its creations from unauthorized reproduction. Although AI adoption has dramatically accelerated since the pandemic, debates about generated content ownership rights are only just beginning to resolve within the government. 

In a recent survey from McKinsey and Company, the complex interplay between generative AI and intellectual property (IP) rights emerged as a significant concern for businesses. Notably, 46% of survey respondents identified IP infringement as a major business risk arising from their deployment of generative technologies. This risk was so pertinent that it ranked as the third most flagged concern about the integration of AI in the workplace among businesses that already leverage generative AI. While the acknowledgment of this challenge is certainly noteworthy, what is particularly disconcerting is the gap between recognition and preparedness: a mere 25% of respondents indicated that they have instituted a formal policy to mitigate the risks of IP infringement related to their use of generative AI.

This chasm between awareness and action reveals a critical blind spot in contemporary business strategy — one that is exacerbated by uncertainty around generative content and IP rights in the legal community. Companies appear cognizant of the unprecedented challenges posed by AI in the intellectual property domain; however, a majority seem unprepared to navigate this evolving landscape. This disconnect not only exposes organizations to potential legal ramifications but also raises questions about ethical stewardship in the age of machine-generated content. It underscores the urgent need for an interdisciplinary approach, combining technological insights with legal acumen, to craft robust policies and frameworks that can guide businesses through the intricacies of IP in the era of generative AI.

The pattern of rapid technological disruption and gradual legal consensus is hardly a new phenomenon. At the dawn of photography, the camera raised profound questions about the nature of art, creation, and ownership. Just as the brush and canvas had been mediums of the artist, the camera became an instrument of the photographer. Yet, there was fierce debate in the legal community: Could a moment captured by a mechanical device, albeit through the photographer’s perspective, truly be “created” and hence copyrighted? This uncertainty mirrors today’s discourse around generative AI. Just as the camera’s role in mediating reality was questioned, so too is the role of AI in producing art, literature, or music. If a photographer’s vision and framing imbued a photograph with originality deserving of copyright, does the programmer’s design or the user’s input in AI systems merit similar protection? Both scenarios challenge our fundamental understanding of creation, authorship, and the rights that should be accorded to those who wield transformative technologies.

Yet the courts seem to differ in their treatment of the two technologies. In the early days of photography, camera-wielders won major victories in front of the Supreme Court in cases like Oscar F. Burrow-Giles Lithographic Co. v. Sarony (1884), in which the court gave a firm admonition of the unauthorized reproduction of the plaintiff’s original photographs on the grounds that a photograph, despite its creation via mechanical means, “no doubt” entails the “original intellectual conceptions of the author.” In a March 2023 Rules and Regulations update from the Federal Register, the US Copyright Office cited Burrow-Giles Lithographic Co. v. Sarony to explain why copyright is contingent on authorship, and why authorship is granted for the photograph and not for generative artwork. 

The Supreme Court, in its interpretation of the term “author” as used in the Constitution, has emphasized the human element in creation. Federal appellate courts have echoed this sentiment, underscoring that those non-human entities, be they animals or machines, cannot be considered authors under the Copyright Act. The case of Naruto v. Slater, which involved a crested macaque named Naruto who took selfies using a photographer’s camera, offers intriguing insights into the ongoing debate surrounding the copyrightability of generative AI products. Just as Naruto could not hold copyright for the selfies he took, generative AI, under current law, lacks the statutory standing to claim copyright for the content it generates. The case serves as a legal precedent that could influence how courts and policymakers approach the increasingly complex landscape of AI and copyright law.

The most pressing question businesses face today is: Who owns AI-generated content? Is it the software developer, given they created the tool? Or the end-user, who inputs the data? Or, in a radical shift, could it be the AI itself? This question is more than philosophical. It bears significant ramifications on licensing, content monetization, and IP infringement.

Additionally, generative AI muddies the waters of originality in copyright law. Historically, originality and fixation were clear benchmarks for copyright. But can a machine’s output, based on vast data inputs, be deemed ‘original’? And if so, who gets credited?

Furthermore, AI’s potential to generate new inventions challenges patent law. How do businesses patent a product that a machine conceptualized? Does the credit go to the machine, the developer, or the operator?

In the burgeoning field of artificial intelligence, patentability has become a contentious issue, manifesting a host of patent challenges that both innovators and legal experts must grapple with. Traditionally, patents have served as the cornerstone of encouraging human ingenuity, granting inventors exclusive rights to their breakthroughs. However, the advent of AI, particularly generative algorithms capable of designing new products or optimizing processes, disrupts this long-standing paradigm.

The legal challenges in patenting AI-generated content revolve around the definition of “inventor,” the ambiguity in the conception of the invention, and the evolving role of AI in the invention process. Policymakers are grappling with these complexities to balance the rights of human inventors and the potential for AI to contribute to innovation.

The current U.S. patent law stipulates that only a “natural person” can be an inventor, thereby excluding AI systems from this designation. This legal framework was upheld in the 2022 case Thaler v. Vidal, where the court ruled that a machine could not be listed as an inventor. The court affirmed that the term “individual” refers exclusively to human beings. This ruling has significant implications for businesses that rely on AI for innovation, as it clarifies that only natural persons can be inventors under the Patent Act.

As businesses integrate artificial intelligence into their branding strategies, a labyrinth of trademark issues emerges that challenges the conventional legal frameworks. Traditionally, trademarks have served as the unique identifiers of a brand’s products or services, but what happens when an AI-generated logo or catchphrase becomes the face of the brand? This question introduces complex trademark protection dilemmas that are just beginning to be explored by legal scholars and industry leaders. 

In stark contrast to the complex legal considerations surrounding projects co-created by humans and AI, the realm of trademarks offers a more straightforward path. Unlike patents or copyrights, trademarks do not necessitate human authorship for protection. A trademark can encompass any word, name, symbol, or device—or a combination thereof—and its eligibility for registration hinges on established legal criteria such as “use in commerce” or “intent to use,” rather than the origin of its creation. It must also meet the criteria of not being generic and being capable of differentiating the applicant’s goods or services from those of competitors. Given this advantage, AI will likely become an invaluable tool for businesses seeking to brainstorm potential trademarks. 

On the other end of the trademark spectrum, businesses that seek to leverage AI tools to create trademarks of their own will still run the risk of infringing on the trademarks of other firms. If the output of the user’s prompt is deemed too similar to an existing brand, there’s no guarantee that the United States Patent and Trademark Office would look the other way. Training datasets often make use of trademarked content, and there are pending legal cases in state courts (such as Getty Images, Inc v Stability AI, Inc) that will eventually provide more detailed information about the court system’s definition of a trademark infringement in the generative context.

Recent legal confrontations over copyrights for AI generated works provide some insight into the unfolding legal precedent that may come to define intellectual property in the generative context  more broadly. Early generative AI legal cases center on the concepts of authorship, fair use doctrine, and the case-by-case nuances of each work presented to the US Copyright Office for consideration. 

The case of “Zarya of the Dawn,” a graphic novel that employed generative AI in its creation, serves as a compelling touchpoint in the ongoing discourse surrounding AI and copyright. The U.S. Copyright Office granted limited copyright protection to the text and arrangement of images in the novel but denied protection for the individual AI-generated images. This decision has been viewed as a partial victory by the creator’s legal team, who argue that the focus should be on the creative input of the human user rather than the output of the AI.

The novel employed a form of generative AI known as MidJourney to produce a series of images. The Copyright Office initiated a secondary review upon learning that the images were produced using this AI tool. The crux of the inquiry was whether the creator’s interactions with MidJourney were sufficient to constitute an independent, creative work. The Copyright Office applied the Feist test, which looks for an independent creation with a certain minimum amount of creativity before granting a copyright. In this case, the Copyright Office identified too much “distance” between the user’s input and the AI’s output to warrant copyright protection.

The decision has broader implications for other generative AI platforms, including ChatGPT and DALL-E. These platforms can expect similar treatment regarding their AI-assisted content generation. The key question that remains is: How much control must a human user exercise over generative AI to receive a copyright for the output? 

Generative AI platforms also face legal challenges related to the use of images for training without a license and the risk of infringing upon protected work. These challenges, coupled with varying legal and regulatory requirements globally, make the future of copyright a complex and evolving landscape.

The intersection of generative AI and intellectual property rights is a fascinating place to be, and the legal and business communities find themselves at a crossroads. The intricate dance between technological innovation and legal frameworks has reached a turning point, forcing us to confront challenging questions around authorship, originality, and ownership. While generative AI offers tantalizing prospects for streamlining creative processes and innovating new products, it also opens a Pandora’s Box of ethical and legal quandaries that businesses and policymakers are only beginning to unpack.

From the courtroom debates on the copyrightability of a macaque’s selfie to the complex legal terrain of AI-generated novels and artwork, the shifting paradigms of intellectual property demand our acute attention. Traditional definitions of authorship and ownership are being upended, and the gap between legal preparedness and technological advancements continues to widen. Whether it’s the murkiness surrounding the patenting of AI-generated inventions, the evolving landscape of copyright in the context of machine-created art, or the straightforward yet complex domain of AI-generated trademarks, each aspect necessitates a creative legal strategy to avoid emergent compliance pitfalls such as trademark infringement.

Back