By: Soussin Emmie
1) AI vs. IP: A Quick Introduction
With the rapid development of artificial intelligence, particularly generative models like GPT-3, AI is not just a tool, it has become a creator. This evolution raises fundamental questions about authorship of AI-generated works under intellectual property law. This raises some questions regarding can an AI algorithm be considered an inventor or author; and how should our current legal system adapt to this new reality.
“Reason is nothing but reckoning.”[1] With this phrase, the 17th-century philosopher Thomas Hobbes insinuated in his book “Leviathan” a future where human reasoning could be reduced to a set of calculations, which is essentially the core concept behind artificial intelligence today.
While Hobbes was ahead of his time in considering human thought as something mechanical, it was not until the mid-20th century that these philosophical ideas began to take life in concrete technology. One of the pioneers of this field, Alan Turing, introduced the “Turing Machine,” which is considered the foundation of what we know today as AI. This machine was capable of performing cognitive processes through the use of algorithms. Turing also gave his name to the “Turing Test”, that measures a machine’s ability to act like a human.
Fast forward to today, Turing’s vision has come to life in the form of an advanced generative AI systems such as the GPT-3, launched in 2020 by OpenAI. Thanks to deep learning and big data, AI has developed rapidly, particularly with the explosion of generative AI.
This phenomenal expansion of AI has raised many interesting issues, especially regarding its impact on the legal profession and intellectual property laws, including copyright and patent law.
For instance, let’s take a closer look at the famous case of Device for the Autonomous Bootstrapping of Unified Sentience (“DABUS”)[2], an AI algorithm. This case concerns two patents’ applications and a copyright’s application filed in 2018 in the US by Dr. Thaler. Thaler claimed that DABUS created inventions and a novel artwork without human involvement. This gives rise to a critical question: Can a machine be an inventor? In this case, the court answered no. Subsequently, the US Copyright Office denied Thaler’s copyright applications for one critical reason, there was no human authorship.[3]
But let’s dive deeper into the heart of the “authorship” issue.
2) Who Is the Author?
Music compositions, visual arts, literature, photographs, poems… AI has become an incredible tool capable of mimicking almost all forms of human art and creativity.
But for now, the main question is: “Who is the author?”
Under the U.S. Copyright Act, exclusive rights are granted to the “author” of an original work of authorship once it is fixed in a tangible medium of expression.[4] Although the statute does not explicitly state that authors must be human, the U.S. Copyright Office and courts have consistently reinforced the human authorship requirement, which is also consistent with the DABUS case.
Similarly, U.S. patent law also necessitates that inventors must be natural persons. According to 35 U.S.C. § 100(f), an inventor is defined as an “individual.” Courts have interpreted “individual” as a natural person, therefore excluding AI and other non-human entities from being recognized as inventors. The United States Court of Appeals for the Federal Circuit explains this idea in the Thaler case. The Thaler court stated that “[t]he Patent Act requires that inventors must be natural persons; that is, human beings . . . . The Act similarly defines ‘joint inventor’ and ‘coinventor’ as ‘any one of the individuals who invented or discovered the subject matter of a joint invention.’”
These statutory provisions were at the heart of the DABUS patent case, where the U.S. Patent and Trademark Office (the “USPTO”) rejected AI as an inventor. The USPTO’s decision highlighted that only human beings can be recognized as inventors. Consequently, both the courts and IP offices have repeatedly rejected claims for copyright and patent protection where the purported creator or inventor is an AI algorithm, maintaining the status quo in favor of human authorship and invention.
On March 17, 2023, Thaler filed a petition with the U.S. Supreme Court to clarify whether the US Patent Act restricted the definition of “inventor” to only human beings. The Supreme Court refused his petition.
As we can see, the US, like Europe, is adhering to a traditional interpretation of texts adopted at a time when AI did not exist.[5] Today, we are undergoing a real technological revolution, especially with advanced machine learning, which is an algorithm fed with data that is able to autonomously produce output and even to learn new things by itself. Consequently, if AI can independently produce results without any substantive input from a natural person, should we still limit the status of “inventor” or “author” only to human beings? This question becomes particularly relevant when neither the original programmer, the AI’s owner, nor its user has made a material contribution to the invention or authorship process.
In the meantime, we’re still in a gray area. Sofia Vescovo, in the “Rise of the Machines: The Future of Intellectual Property Rights in the Age of Artificial Intelligence,” explains that some legal scholars have begun proposing new approaches, such as “Digiworks Rights.” The idea is to create a new rights category; AI may not fit into the “human author” category, but a new subcategory could be established. This new category would be called “Digiwork,” where the source of the work is Artificial Intelligence. [6]
For now, the U.S. Congress and IP offices appear to prefer maintaining the status quo. But change seems to be only a matter of time.
In a similarly intriguing case regarding the classification of types of authors, it’s not just machines that have been targeted, it’s animals too. An example is the Naruto v. Slater case, often referred to as the “Monkey Selfies” case. In the Monkey Selfie case, People for the Ethical Treatment of Animals (the “PETA”) argued on behalf of Naruto, a 6-year-old macaque, claiming that the monkey was the author and copyright holder of the photos at issue. However, the US Court of Appeals dismissed the complaint, ruling that monkeys are not humans and therefore lack statutory standing under the Copyright Act.[7]
We now know that AI, like a monkey, can’t apply for copyright or patent protection yet, but AI has many other capabilities. Not only can it create art, but it can also revolutionize how work is done.
3) Is AI a Job-Killer?
If I were you, I wouldn’t become a lawyer, AI is going to steal your job, and you’ll soon be useless …. I appreciate the concern, but I prefer to think of AI as a helpful colleague, not a job thief. I strongly believe AI will be a valuable tool to help practicing law, rather than replace our work. This is especially true in the field of Intellectual Property, where AI is already transforming work processes and creating new opportunities.
To understand this fear, let’s do a quick comparison to some earlier technologies. When spreadsheet software like Excel, VisiCalc, and Lotus 1-2-3 first emerged, concerns arose that these programs could replace accountants. These fears arose from the automation of complex calculations, data entry, and financial analysis, which were previously done manually. However, those technologies turned out to be a transformation, not a replacement. The accountant’s role evolved, efficiency increased, and new jobs emerged, such as data management or software specialization. Ironically, today, expertise in Excel is considered essential for accountants.[8]
In the IP world, a similar trend is emerging. Paximal, an AI technology launched in 2023, automates IP processes, including the auto-generation of patent documents. It was co-founded by Ian Shick and Michael Drapkin. This tool can analyze your patent application and draft a patent in a fraction of the time. As a result, what once took days can now be done in hours, and of course the cost of this process would also be dramatically reduced.
Naturally, this raises the question: Will this reduce job opportunities for patent lawyers?
The answer remains unclear, but it seems that such a process will make it easier to apply for protection of inventions, reducing both cost and time. This means more inventions, more applications, and, hopefully, even more work for lawyers. It could also allow lawyers to focus more on the claims themselves. Therefore, we can expect AI to change the nature of a patent lawyer’s work by automating repetitive tasks and improving research efficiency.
According to François-Xavier Leduc, CEO and Co-Founder of DaVinci/Kili Technology, “[t]he strategic opportunity lies in saving 50% of attorneys’ time today and envisioning a future where the arduous task of writing a 40-hour patent could be accomplished in just four hours, totally redefining the landscape of patent writing and prosecution.”[9]
One interesting point to consider is that AI may not replace patent attorneys, but patent attorneys who use AI will likely replace those who don’t.
Lastly, knowing that AI will continue to evolve, the legal landscape must also adapt to those evolutions to ensure that future protections will exist. The challenge is in integrating AI into our legal frameworks while preserving the fundamental principles that govern authorship and invention.
In the light of these considerations, artificial intelligence is no longer a subject of science fiction but a reality that demands serious attention. We need to know that we cannot control and stop the growth of AI, so we need to evolve alongside it; boundaries, such as the power that we will give to AI, need to be carefully controlled.
Who knows? Perhaps an AI wrote this paper.
Soussin Emmie.
[1] Thomas Hobbs book: Leviathan
[2] Thaler v. Vidal; herein referred to as the DABUS case or the Thaler Case
[3] https://patentblog.kluweriplaw.com/2024/01/16/the-end-of-the-road-for-dabus-and-dr-thaler-at-the-uk-supreme-court/#:~:text=This%20case%20concerns%20two%20patent,involvement%20of%20a%20human%20inventor.
[4] 17 U.S.C. § 102(a)
[5] https://www.saderlaw.com/news.php?reader=242
[6] Sofia Vescovo, Rise of the Machines: The Future of Intellectual Property Rights in the Age of Artificial Intelligence, 89 Brook. L. Rev. 221 (2023).
[7] https://cdn.ca9.uscourts.gov/datastore/opinions/2018/04/23/16-15469.pdf
[8] https://www.linkedin.com/pulse/why-excel-enemy-accountants-michael-whitmire/
[9] https://ipwatchdog.com/2024/01/12/transforming-ip-practices-rise-ai-powered-patent-attorneys-2/id=171885/