WHO WROTE THIS ANYWAY? COPYRIGHTABILITY OF AI WORK AND REALITY BEHIND AUTHORSHIP

Abstract

This paper examines the question of artificial intelligence's (AI) status as an author under copyright law through the lens of constitutional interpretation theories. Utilizing public meaning originalism, we explore the historical semantics of the term "author" around the years 1787-1789, focusing on whether the term referred exclusively to human creators or could encompass any entity capable of creation. Dictionaries from the era serve as the foundation for arguing from an originalist perspective, contrasting this view with the living constitutionalist approach that considers the evolution of language and societal norms. This analysis extends to the examination of historical copyright statutes and relevant case law, including Thaler’s case and instances involving AI, to infer legislative and judicial trends. The argument also encompasses institutional capacity considerations, debating whether the judiciary, Congress, or the Copyright Office is best suited to address this complex issue. Employing constitutional pluralism, the paper synthesizes various interpretive methods to arrive at a comprehensive judgment. Ultimately, this paper seeks to elucidate whether the contemporary understanding of "author" aligns with historical interpretations, or whether our understanding of “author” should be defined with more than a mere meaning of the author at the creation of the Constitution.

INTRODUCTION

Artificial Intelligence (AI) has woven itself into the fabric of our daily existence, becoming a ubiquitous presence that shapes how we interact with the world. From its conceptual origins in the musings of ancient philosophers who pondered automated machines and logic, AI has evolved dramatically. Roots of the artificial intelligence(AI) go way back to the ’50s. As depicted in the famous movie “Imitation Game”, a movie with a historical concept explaining how Alan Turing created the Turing machine, we see traces of artificial intelligence nearly 100 years later after the invention of computers.[1] The first use of the term “artificial intelligence” occurred in 1956 by John McCarthy at the Dartmouth Conference.[2] Improvements in computers and machines allowed us to take a glance at the possible potential of AI. Especially when the computer met the internet, the available data, and the capabilities of these machines beyond our imagination. As we come to the millennium, the introduction of deep learning and neural networks constructed the fundamentals of AI as we know it today. Especially 2010’s were the years of development. This innovation gave the fruit in 2020 and almost all of us are knowingly interacting with AI on a daily basis. AI improved in a way and speed that no scientist had envisaged at the time. Currently, we can use AI systems to have a chat. This is not just a limited talk with old chatbots. AI systems are trained by a vast amount of data. Therefore, they have the ability to answer more profound questions rather than trivial daily talk which will make them an unassailable part of work life in the near future. Moreover, it is possible to create images with AI. Considering the fact that most of the pictures used online are copyrighted, using AI image creation for projects or daily use brings no copyright issues because of current copyright law. Finally, AI can make music now. Just like with the images, music generated by AI is also not protected by copyright laws. But are they public domain? Does this mean that copyright protection will become less important? Doesn’t this approach deter authors from progressing science? This rapid enhancement brought legal discussions and disputes need to be resolved. Today, legal issues regarding AI turning into a spectrum from its personhood to its liability. The question that started to be seen lately is the patentability of its creations. Can AI be an inventor? In order to answer this question, we must inspect and analyze the article in the Constitution that granted power to Congress to enact a statute to promote sciences and useful arts.

SECTION 8 OF THE CONSTITUTION—INCEPTION OF ALL

U.S. Constitution included an article about the protection of the products of authors and inventors by giving power to Congress to enact such a statute. Article 8 of Section 8 of U.S. Constitution states;

“The Congress shall have Power . . . To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries;”

As seen, the framers of the Constitution wanted to protect authors and inventors to keep science and useful arts progressing. The best approach to reach a sound conclusion would be using the approaches from different constitutional views. First, we should use “public meaning” originalism which means, what reasonable and informed members of the public would have ascribed to the text at the time of its promulgation.[3] To utilize public meaning originalism, we should look at the meaning of “author” and “inventor” for our purposes. According to the Merriam-Webster Dictionary, the first meaning shown for the “author” is “the writer of a literary work.”[4] In order to understand its use, we should look at the use between 1787 and 1789. Oxford English Dictionary provides historical meanings of the word. One that is very interesting and still in use today is “the creator of nature, the universe”[5]. This meaning shows us that the word is not only used for human beings. Therefore, at the time the Constitution was amended, “author” had more than a plain meaning. Regarding the “inventor”, the meaning used is “One who devises or produces something new (as an instrument, an art, etc.) by original contrivance”.[6] This also points out that an inventor can be anyone producing something new. Thus, although there is a strong possibility that framers only thought about human beings, from the public-meaning originalist perspective, we cannot eliminate the possibility of “author” and “inventor” being someone other than natural persons. If we approach from the living constitution perspective, we should consider how big AI takes place in our lives. Since the living constitution view involves the idea that the constitution should adapt to changes in social life and the nation’s circumstances, AI’s ubiquity and importance in art and science have become an important part of the answer.[7] Since the commencement of ChatGPT of OpenAI, people started to realize that AI improved beyond their apprehension. Although they are not used by everyone, a lot of people use ChatGpt or similar AI bots in their daily lives. Especially considering the fact that OpenAI will release a ChatGPT-5, and tech firms training large language models for specific areas, Harvey training a huge LLM for law firms, there is no hesitancy that in the next 5 years, AI chatbots will be an inseparable part of our lives.[8] Because AI will be indivisible for us, the living constitution perspective will be inclined to the idea that AI can be an author. This understanding does not find its basis solely from social norms. Social norms are just affecting the disposition of the living constitutionalist. Since an author can be used for things that are not human, at least the idea of AI being an author can be supported by this approach. The living constitutionalist may support its view by deep diving into the words. The contemporary meaning of the author is “the writer of a book or other work; a person whose occupation is writing books.”[9] Since it does not specify whether the author will be a human or not for the first part, the living constitutionalist may claim that contemporary meaning of the author changed. However, looking at the sole wording of the Constitution is not enough. We should move forward to see what the actual intent of Congress is. In light of the clause mentioned supra, Congress enacted the Copyright Act in 1790. The Act changed throughout time and we should inspect the changes and meanings of the provisions.

THE COPYRIGHT ACT

The first Copyright Act was narrowly drawn, and only mentioned several products. The text read as follows;

“That from and after the passing of this act, the author and authors of any map, chart, book or books already printed within these United States, being a citizen or citizens thereof, or resident within the same, his or their executors, administrators or assigns, who halt or have not transferred to any other person the copyright of such map, chart, book or books, share or shares thereof…”[10]

Congress granted protection to authors on maps, charts, and books. The understanding of Congress for “author” was obviously a human being. The text mentions authors being citizens and residents which are attributes of a natural person.

The Copyright Act had big revisions in 1831 and 1909. The last amendment to the Copyright Act was in 2022, which amended the Copyright Act of 1976. Today, the Act contains a definition section. Although there is no direct definition of author, the text mentions authors on several occasions. Most importantly, U.S.C. Title 17, §101 states that “anonymous work” is a work on the copies or phonorecords of which no natural person is identified as author. Text in the provision implies that the author can only be a natural person. However, the Provision that is mostly subject to the cases is a different one. U.S.C. Title 35, §100 also has a definitions section that expressly states the meaning of “inventor”. The text says;

“The term “inventor” means the individual or, if a joint invention, the individuals collectively who invented or discovered the subject matter of the invention.”

Although the definition of the individual is not given, the Supreme Court explained that when used as a noun the word “individual” means human beings, a person. Mohamad v. Palestinian Auth., 566 U.S. 449, 454, 132 S.Ct. 1702, 182 L.Ed.2d 720 (2012). Therefore, the view of the law itself is not in favor of AI. For AI protection, the Copyright Act needs an amendment for the reasons explained in forthcoming sections.

COPYRIGHT OFFICE’S APPROACH—LATEST GUIDANCE

In early 2023, the Copyright Office published new guidance about AI work. The reasonings of the Copyright Office will be included in different sections of this article. Most of the reasoning behind the guidance is about the “human authorship” requirement and the courts reasoning related to that. Guidance includes a term called “traditional elements of authorship”[11] Especially, the Office focuses on products generated with prompts to AI and states that these products lack human authorship requirement. Office considers the process of prompting as merely instruction, therefore claims that the process itself lacks creativity. By traditional elements of authorship, Office concentrates mostly on the human authorship requirement. The argument is mostly whether the person has the creative control over the product or whether AI has creative control.[12] But based on the examples given in the outline, Office does not want to approve an authorship regarding AI even slightly. In the example, Office states if the author creates a work by arranging and selecting AI-generated images, the work will constitute original work of authorship but only the human-authored parts will be subject to copyright protection.[13] Therefore it is obvious that Office pays significant attention to the human author requirement.

CURRENT CASES

Although AI is improving rapidly, cases involving either AI inventions or infringement regarding AI are not common. One possible explanation for this is that people do not state in their patent application that the product is created by an AI.[14] However, despite the fact that there are not many judgments regarding AI, some project cases started to be tried in the courts and outcomes may not be in favor of AI. The 2 cases that stand out involve a scientist and his machine that creates different images. For this discussion, we should turn our direction to one of the most important cases that examine the possibility of autonomous creation and artwork.

Thaler v. Vidal            :

Thaler, a chemist by training, had a near-death experience when he was young. This experience would lead him to invent maybe the first AI-created artwork in the coming years. While trying to invent a computer model that would simulate knock-on damage of atoms for his research projects he came up with a different finding.[15] When he started using the model for fun, he started damaging random photographs which was the beginning of his experience with AI and copyright. He created “Device for the Autonomous Bootstrapping of Unified Sentience”, in short DABUS, which generates variations of different pictures with the input it has. We can see this machine as an AI system trained with a limited amount of data. The important part about Thaler’s approach and invention is he does not see DABUS as a sole machine. He believes he has emotions and deliberately uses the term “sentient machine.” In 2012 he introduced a different kind of noise: simulation of the near-death experience he’d had as a child. He intentionally severed a portion of DABUS neural nodes from the rest of the network and found that it caused a reaction similar to a human’s end-of-life light show, something Thaler calls “life review and then the manufacturing of novel experiences.” Afterward, DABUS began reviewing its data or, as Thaler puts it, its “memories,” and, from them, produced an image showing train tracks threading through brick archways that it called A Recent Entrance to Paradise.[16] Thaler’s perception of AI prompted him to undertake actions on licensing of AI as the author of the products produced by DABUS. Thaler filed his application on behalf of DABUS and mentioned DABUS as the creator and author of the images. As expected, the United States Patent and Trademark Office did not approve this application reasoning that only human beings can be authors. Consequently, Thaler brought civil action on this matter. Federal Circuit 4th made findings on whether an AI software system can be listed as an inventor on a patent application. Thaler has sworn on behalf of DABUS and purported himself all the rights of the DABUS as an inventor. The court considered first whether the language of the text is ambiguous. It found that statutory text is not ambiguous and inventors must be natural persons that are human beings. The reasoning for this finding was The Patent Act’s provision. §100 of the Act states,

“The term “inventor” means the individual or, if a joint invention, the individuals collectively

who invented or discovered the subject matter of the invention.”

Even though the Act does not provide a definition for the word individual, the Supreme Court explained that when used as a noun the word “individual” means human beings, a person. Mohamad v. Palestinian Auth., 566 U.S. 449, 454, 132 S.Ct. 1702, 182 L.Ed.2d 720 (2012). Congress could have used “person” or “whoever” in the definition which might have raised questions on whether other legal entities could be authors too. However, congress specifically worded the provision with “individual.” It is blatant that Congress only wanted to include human beings as authors. The court also mentioned that the Act uses “himself” or “herself” to refer to an “individual.” Court’s interpretation of the “author” starts and ends with the plain meaning of the text since the text itself is not ambiguous.[17] Thaler also makes arguments about the canon of constitutional avoidance, however, the court answers these arguments by explaining the relevant article of the provision is about the power granted to Congress, and Congress chose to exercise this power by limiting the requirement of being an author to human beings.

Although Thaler could not get the result he wanted, he was a party to another civil matter regarding DABUS.

Thaler v. Perlmutter                :

Just like the case above, Thaler sought the registration of the artwork created by the DABUS. Since he was the user, owner, and developer of the Creativity Machine, he claimed he was entitled to its rights.  Of course, the Copyright Office denied the registration because the work lacked a human author.[18] However, Thaler was not the only one testing the system. Kristina Kashtanova also tested the system by trying to register a comic book that included images generated by AI.[19] Although the copyright office registered the product in the first place, after the announcement of Kristina Kashtanova on AI-generated image registration, the Copyright Office informed her that they were considering rescinding the registration. They concluded that the book written by Kristina Kashtanova herself has protection under the Copyright Act. However, the statute does not protect the images created by the AI. Just like Kristina, Thaler also received a similar result in the judgment he sought against the Copyright Office. U.S. District Court for D.C. granted summary judgment for the Copyright Office by the following reasoning;

Copyright is designed to adapt with the times. Underlying that adaptability, however, has been a consistent understanding that human creativity is the sine qua non at the core of copyrightability, even as that human creativity is channeled through new tools or into new media .... Nonhuman actors need no incentivization with the promise of exclusive rights under United States law, and copyright was therefore not designed to reach them. The understanding that “authorship” is synonymous with human creation has persisted even as the copyright law has otherwise evolved.[20]

Although the case is on appeal in the U.S. Court of Appeals for the D.C. Cir., courts seem to maintain an averse view on the copyrightability of AI products. The court also observes other cases that are deemed as cases involving non-human work. One of them is Urantia Found. v. Kristen Maaherra, 114 F.3d 955, 958–59 (9th Cir. 1997). Urantia case involves a book that is claimed to be written by a spiritual. According to facts, it is just physically written by a person. Although the product is supposedly the work of a spiritual creature, the court decided that the human involvement in terms of writing the book physically is enough for copyrightability.  Now we will change our direction to a case where the anthropocentric view of the Copyright Act is discussed. However, unlike the other 2 cases we examined, this case questions the human authorship with something other than AI—a monkey.

Naruto v. Slater           :

Slater is a wildlife photographer. The action consists of claims brought by a monkey. Yes, you read it correctly. People for the Ethical Treatment of Animals (PETA), brought a claim on behalf of a Crested Macaque monkey—Naruto.[21] Naruto was a seven-year-old monkey who lived on a reserved island where the photographer left his camera unattended.[22] Monkey took selfies of himself. These pictures are published in a book Slater created and he identified himself as the copyright owner of the monkey selfies.[23] These events unfolded in the case of Naruto v. Slater. Even though the court was not satisfied with the “next friend” representation of PETA and found it insufficient, the court delved into consideration under The Copyright Act because the court believes that it is obligated to consider whether incompetent parties are adequately protected.[24] The court follows its precedent for the ruling. According to the precedent, the court did not rely on the fact of whether the statute mentions them as “persons” or “individuals”. The court simply said;

“If an Act of Congress plainly states that animals have statutory standing, then animals have statutory standing. If the statute does not so plainly state, then animals do not have statutory standing. The Copyright Act does not expressly authorize animals to file copyright infringement suits under the statute.”

The court established this rule in Cetacean Cmty. v. Bush, 386 F.3d 1169 (9th Cir. 2004). Even though the court used this approach for statutory interpretation, Thaler v. Vidal, explained supra, examined the language in the statute and it was not in favor of non-humans. The court also elucidates the statutory text in a similar way to what the 4th Circuit did in Thaler v. Vidal. The court said;

“The terms “children,” “grandchildren,” “legitimate,” “widow,” and “widower” all imply humanity and necessarily exclude animals that do not marry and do not have heirs entitled to property by law.”[25]

Therefore, taking into account the cases mentioned earlier, it’s evident that the path of copyrightability is not only closed for AI but also for non-human entities. Although the court entered a judgment for Slater, it did not answer important questions regarding the issue. It is not explained whether Slater can enforce his own copyright on the photos.[26]

Burrow-Giles Lithographic Co. v. Sarony      :

Although this case goes way back to 1884, it is a noteworthy case to present an action taken by the Supreme Court to adapt to technology. The plaintiff took the photograph of Oscar Wilde by making all the necessary arrangements involving the steps to take the photograph.[27] Burrow-Giles Lithographic Co. uses the photograph without the permission of Sarony. The claim asserted by Burrow-Giles was that the photograph is just a capture of an existent scene and it has no creative aspect. According to view, since it is not a writing or creative product it is not copyrightable. The Supreme Court highlights the amendments in the Copyright Act. Just like explained below, the Court stated that the Act changed over time, and new sections were added to protect works other than writing such as paintings and engravings. The Court shows the pattern of improvement of logical reasoning and therefore expansion on the statute’s range of protection. The Court found that Sarony carefully orchestrated the scene that was captured, and Congress and the Constitution intended to protect this work.[28] Therefore, the Court ruled that copyright protection for Sarony’s work is appropriate. Even though this case is not about the human requirement of the Copyright Act, it shows the adaptation of Courts to technological developments. With this regard, we should also examine the other requirements for a work to be copyrightable. These requirements are not really an obstacle in the way of AI. Most of the work produced by AI can satisfy the requirement.

In order to be copyrightable a project must be “original works of authorship fixed in any tangible medium of expression.”  17 U.S.C. § 102(a). The key requirements here are originality and fixation. [29] These requirements can be comprehended as the traditional elements of authorship, the term used by the Copyright Office lately.

TRADITIONAL ELEMENTS OF AUTHORSHIP

For the last 2 years, The Copyright Office has used a term for the forthcoming elements of authorship. Office mentioned “traditional elements of authorship” which include human authorship and causing applications using AI to be rejected. Office took a very aggressive position against AI work, any work that passes beyond de minimis standards got rejected in part or as a whole.[30]

Originality

A work is considered to be original for copyright purposes if;

• it was created by the author, rather than copied from other works and

• it possesses at least some minimal degree of creativity.

Feist Publ'ns, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340, 345 (1991).

The bar for creativity is significantly low and it is mostly easy to meet the requirement for AI products. While we may discuss the rejection of AI’s creativity on the grounds that it is not capable of creating or thinking and is just a mere production of the inputs uploaded into the system, we can make a comparison between humans. Humans also use their knowledge and experience in terms of creativity. As the ancient philosophical axiom says, “ex nihilo nihil fit” which translates to “nothing comes from nothing,” it is impossible to create something without having an input in technological terms. Therefore, AI can satisfy the originality prong.

Fixation

The product is thought to be “fixed” if it is in a tangible medium that is sufficiently permanent or stable to permit it to be perceived, reproduced, or otherwise communicated for a period of more than transitory duration.[31] It may be perceived, reproduced, or communicated directly by a person or with the aid of a machine, such as a computer. 17 U.S.C. § 102(a). Therefore, most of the obstacles to the copyrightability of AI work derive from the human authorship requirement.

Human authorship

As we observed above, human authorship is the predominant impediment to the copyrightability of AI work. Despite the fact that it is a barrier, it does not hinder people from registering the products of AI. Research shows that AI-generated works likely have been registered with the Copyright Office despite this requirement. People can bypass this requirement in several ways. One could simply file an AI-generated work and list a person as the author of the product.[32] Moreover, it is possible to file for an AI-generated work as Work Made for Hire (WMFH) if there is no requirement to list an author.[33]

WHY HUMAN AUTHOR

There are different reasons for the human authorship requirement. The most fundamental one is the idea that machines can never be like humans.[34] The belief is that since AI does not have a conscience and understanding of the world and knowledge as we perceive it, it can barely imitate the human mind.[35] The main argument to raise here is we are not fully sure how we humans make our judgments. It is obvious that some form of emotions and understanding is involved. Does our sentiment really play a role in everything we do? In other words, feelings may have an effect on judgments or conclusions, but not every step we take involves them.

Another argument is even though humans and AI can create the exact same work, the process is essentially different for a human than for an AI.[36] Therefore it is claimed that it can never be the same thing. This argument behind the requirement is just an idea. There is no scientific or social background to this argument.[37] Since we are not sure how the creativity of the human mind works, it is not possible to make strong conclusions.

In the end, does the process of the work really matter? Copyright is granted regardless of the way the work is made. Therefore, this reasoning behind human authorship does not have any effect on the issue.

JOINT AUTHORSHIP—CAN AI BE PART OF THE PROCESS?

Joint authorship exists when the work has been done by more than one person. In this case, patent applications or copyrights belong to authors. One of the cases that structures the understanding of Joint authorship is Aalmuhammed v. Lee, 202 F.3d 1227 (9th Cir. 2000). In the case, Aalmuhammed was a technical consultant on a movie produced. The movie was about the life of Malcolm X, co-produced and co-written by Spike Lee, and starred Denzel Washington.[38] During the making of the movie, Aalmuhammed helped producers to depict the life of Malcolm X as closely as possible. He was knowledgeable about his life and Islam, so he was the consultant during the process. Subsequently, he was listed in the credits section as an “Islamic Technical Consultant”.[39] He filed the complaint claiming that he helped with the movie, and he is one of the authors. Therefore, he is one of the copyright owners. The biggest takeaway from the Aalmuhammed case is that although some people may be involved in the process of the work, that does not necessarily mean that they are authors. [40] The court stated that the author is used as the originator of the work. It is further mentioned that during the process of movie making, a lot of people are involved such as producers, editors, animators, composers of the music, etc. In the eyes of the law, it is not feasible, possible and just to accept everyone involved as authors. Although in the Feist Publications, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340, 111 S. Ct. 1282, 113 L. Ed. 2d 358 (1991) court stated that more than minimal originality and contribution is enough, in the context of movies, the author and mastermind behind all would be the one with “artistic control”[41]

THEORIES THAT LAWYERS CONSIDER

There are 5 common theories for authorship in AI work. Firstly and current approach is there is no author on AI work. Secondly, the creator may be considered an author. Thirdly, the user of AI as a tool is the author. Fourthly, The user and creator of AI are joint authors, which will be mentioned in the next heading. Finally, the doctrine of “Work Made For Hire”.[42]

1.AI Work Do Not Have An Author

This view derives from the current law. Since there is no rule allowing copyright to AI, work made by AI can simply be considered public domain. As we mentioned before, current law allows copyright to only human beings. Therefore, a process that lacks a human would not be protected under copyright law. This resembles the Naruto case, supra. Naruto was also rejected for copyright because of lacking the human capacity. This view is justified with an economic perspective by some lawyers. With the vast use of AI, some people started utilizing AI just to have economic benefits. They try to get a product out of it, based on solely financial purposes. Blatantly, this contradicts the main purpose of the intellectual property clause in the Constitution. The intellectual property clause aims for the progress of science by giving exclusive rights to authors for a limited time. However, the economic benefit being the sole purpose, might endanger this goal by not progressing science in a meaningful way. Although some lawyers believe that there might be important works staying private due to current laws, people can still receive the copyright by showing themselves as authors. Therefore, this public domain view may not be the best approach, but it also does not preclude important works from coming to light.

2.Developer Of AI As Author

This approach finds the reasoning from a comparison of tools. The comparison is between a camera and an AI. A camera is just a tool that would help you to take photographs. However, the creation of the scene, light, contrast, and all the creative parts belong to the person who takes the picture. On the other hand, AI has a different level of skills. It can produce music, write articles, or create images.[43] At first glance, the fundamental idea behind this theory seems reasonable. Even so, I would like to emphasize the case we talked about in the sections before–Burrows-Gilles. At the time of the case, people had their initial encounters with cameras. Their view was that the camera is a tool and the scene already exists in the world itself, therefore there is no originality for copyrightability. With more use of cameras, people started to realize that it is not that simple. Now, if we apply this to our views on AI, cameras and AI may not be that different. Firstly, we don’t know how vastly AI will be used during the production of a work. Secondly, if AI is the sole creator, that should mean that everyone can create meaningful work. If that’s the case, why not a lot of people create beautiful music, or images using it? Finally, we do not know which kind of products will be produced by the use of AI in the future. Therefore, it is too early to make a conclusion based on this theory. Some lawyers connect their reasoning with common law doctrines. The law of capture will be the topic of discussion here. According to the law of capture, whoever reduces the thing to her dominion and control would be the owner of the thing.[44] A fundamental example of this is capturing animals.  A hunter can own a wild fox when it is removed from its natural state by trapping or killing.[45] The law of capture is used in the most unexpected circumstances because of its versatility.[46] Furthermore, another example of the law of capture theory is extracting oil.[47] As disputes arose on the ownership of oil on private land of people, many states took the approach on whoever extracts the oil from the well, or the gas from the deposit would be the owner of the oil.[48] Application of this rule in the context of AI work, the owner/developer of the AI would be the owner of the product since they would be the one who draws the work. However, this approach brings more questions than answering the existing ones. What happens when the AI is not a private one and something like a ChatGPT? The user of the AI can also extract a work from it by using the prompts. Would that mean that work belongs to the developer of the AI?

3.The Use of the AI As a Tool Is The Author

This is the most rational theory among the existing theories. It claims that AI is just a tool, and just like in the camera-photographer relationship, the user of AI should be the author of the work. Regarding the originality requirement, we mentioned several times that the bar for originality is really minimal. Feeding AI with prompts and trying to produce something meaningful would satisfy the requirement. As an illustration, it is safe to say that a user of AI who is trying to create a work, will try different and creative prompts, and inputs over and over again to reach something desired. Furthermore, a photographer who happened to be on the scene and took a picture of an event will also satisfy the requirement. Therefore, the level of originality and creativity is not high, it is almost trivial.

Moreover, Professor Jane Ginsburg also gives examples of how creativity is not fully a human touch and there are also other factors. She explains some unintended acts of creation such as images generated by bad eyesight, claps of thunder, and frustrated flinging of sponges.[49] Although these are not the author’s own actions, the author can adopt a product that is derived from these.

Despite the fact that it seems reasonable, some of the lawyers still contest this view. The reason is, that it is not easy to accept someone as an author, who just typed some words and clicked the button. Repeating again, if this is the point of contention it is hard to imagine those lawyers at the time of Burrows—Gilles.

Outside of copyright law, common law might also have a solution here. We mentioned the law of capture above. As we consider extracting the work like capturing an animal that is in a natural state, we might catch a resemblance. We can claim that work prompted through AI was in a natural state in AI’s database.[50] It was not meaningful or perceivable. We drew the work from a natural state with the prompts and now have ownership of the work. It resembles a fisherman catching a fish and having ownership.[51]

Both the comparative view on cameras and the law of capture doctrines show that the author of the work can be the user of the AI. Although this answers the question of a work that has human intervention, it might not be the answer for fully autonomous works that we will see in the future.

4.Developers and Users of AI as Joint Authors

The fourth possible theory is the joint authorship of the developer and user of AI. The joint authors exist when the work of both of the authors is not meaningful as an independent work. This suggests that the work should have no meaning independently and should be inseparable.[52] These requirements for joint authors stay as an obstacle in the way of this theory. Because when it is observed, it is visible that AI has a meaning outside of the work. Since the AI system itself is separable and meaningful without the work, it would not suffice the joint authorship. However, for specific AI systems, it would be arguable that the AI system has no meaning without the work. This possibility is illustrated in Cumberland Law review.“If a developer creates an AI specifically intended to write recaps of college basketball games, it is reasonable to claim that AI achieves its primary significance when it is fed box scores and produces those recaps.”[53] Looking back to Aalmuhammed v. Lee, the deciding factor in authorship would be who has the creative control over the work. Although it is disputable, AI seems the one standing out. No matter who has the creative control, it is clear that this theory fails due to this requirement.[54]

5.Works Made For Hire

The final theory for AI work is the product belongs to the user under the works made for hire doctrine. The doctrine is explained in 17 U.S.C §101. It is explained as “ a work prepared by an employee within the scope of his or her employment”. The author of the work is the employer or other person whom the work was prepared for. Based on this doctrine AI would be the employee and the user of the AI would be the employer whom the work was prepared for. Although it seems reasonable, it has not been adopted by the courts so far and it does not seem likely. Instead, courts might prefer for Congress to take action regarding the issue, which will be explained in the institutional capacities section.[55]

REASON BEHIND ALL--PURPOSE OF INTELLECTUAL PROPERTY CLAUSE

While delving into the Constitution’s and statutes’ language and trying to determine whether it is meant to be included in the text, we should not disregard the main purpose behind all, what the framers aimed with the intellectual property clause. Current U.S. Copyright Law derives from the intellectual property clause of the Constitution. The clause included in the Constitution reads as follows;

“to promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries”[56]

Before the ratification of the Constitution, there was no unified patent and copyright law since every state held the power regarding that.[57]

As it is understood from the intellectual property clause, the fundamental reasoning behind the clause is to promote the progress of science and useful arts. In order to do that, the clause aims to protect the work of authors and inventors. Thaler used this fundamental clause to structure his argument in Thaler v. Vidal.[58] While the court did not find the argument logical, it is not a distant reasoning especially if there will be statutory amendments or changes in the views of the courts. The only downside is the lower courts would not delve into this argument, rather if the Supreme Court issues a certiorari this argument might be assessed. Considering the change in the meaning of “income” in Eisner v. Macomber, change in the concept is possible.[59] In this case, The Supreme Court excluded stock dividends from income. Therefore the meaning of “author” is open to discussion.

THE MEANING WE ARE LOOKING FOR—THE CONCEPT OF THE AUTHOR

Before this section, we considered the perspective of The Copyright Act and Constitution on the author. We mentioned the cases reflect the courts’ views of the issue. Courts mostly based their reasoning on the statutory text of The Copyright Act which includes “herself, himself” and the meaning of the “individual”. Because the Supreme Court examined the word “individual” and decided it means humans, courts followed these two reasons in their holdings. However, we should not reach a conclusion hastily, and should further think about our perception of the author. Mere examination of the word at the time of the constitution would not be helpful for the argument. Especially considering that the meaning of the words can change over time. For instance, in the U.S. Constitution, “domestic violence” is used for riots.[60] Today it is being used for violence within the family. Just like that, the word “arms” in the Second Amendment meant muskets and flintlocks in 1791, however today most people bear the latest technology guns like Glock and AR.[61] Moreover, as we observed in previous sections the statutory text of The Copyright Act uses “herself” and “himself” to refer individuals. Courts argued that these individuals who are authors can only be humans because of the gender pronouns. However, at the time of the amendment, there were only 2 accepted genders. Today there are people who are outside of that traditional understanding who identify as non-binary. With this reasoning Copyright Office would not give authorship to those people. It is obvious that this assumption is not true but it is a good indicator that our perception of words can change quickly. Therefore, sticking to the meaning of the framing of the Constitution or other text is not the right approach for this problem.  We should examine the concept of authorship. What kind of attributes do we imagine when we say author? We are not asking ourselves who are the authors. Rather we think about a prototypical author. This author has the ability to create artwork. It creates a meaningful, fixed, and original artwork. Is it really required to be a person to carry these characteristics of authorship? What if non-humans can also carry these attributes in their work? For instance, let’s say we discovered an indigenous people in a remote area, still living a life close to first humans. Assume that they are not homo sapiens in terms of species. If they have some drawings on their cave walls or some sculptures, they would surely be considered authors. Let’s take a step further in this thought process and think about aliens. If aliens start to communicate with us one day and come up with their work, would it be copyrightable? It will probably be because although they would be a different spicy, they will satisfy the traditional elements of authorship. Therefore, it would not be reasonable to reject the work as original authorship just because it does not satisfy human authorship. Just like that, we are facing a different spicy. It is not a living thing as we apprehend but it also satisfies the requirements of the original work. Because of that, we should not reject AI work because the term author we think is a functional kind that’s helping the progress of science and artwork.

INSTITUTIONAL CAPACITIES

One of the questions regarding the copyrightability of AI work is which institution should authorize this. This question includes diverse views, which are supported by a different idea. The first one that comes to mind would be that Congress should take a step to allow it. Considering the fact that Congress’ main purpose is to reflect the public’s view and represent people, it is reasonable to assume that the first step should be taken by Congress. The way it can be done is simple in its face. Congress can amend The Copyright Act by adding a clause about AI. Secondly, the Copyright Office can change its practice and take steps to allow AI’s authorship. Since copyright protections are granted by this institution, grating AI could be a step taken by the Copyright Office. However, the Copyright Office is limited by the statute’s language. Therefore, it is not a practice that occurs to act outside of the language of the statute that regulates the area. Finally, the Supreme Court can also grant copyrightability by changing the interpretation of some of the meanings in the statute. The biggest example of that would be the interpretation of the word “individual”. By enlarging the meaning of the word, because of the changes in technology and society, the Supreme Court can reach a conclusion favorable to AI.

POSSIBLE STEPS TO TAKE TOWARD SOLUTION

In light of all these arguments, it is evident that we need an amendment in the field regarding the copyrightability of AI. After discussing the current situation and institutional capacities involved in copyright law, it is time to gather the information for possible steps that might be taken to address the issue of the copyrightability of AI. First and foremost, as a temporary solution Copyright Office can use the rule of doubt for works of AI.[62] Rule of doubt suggests that if the Office has not taken a position on a legal issue that is directly relevant to the question of whether the work constitutes copyrightable.[63] Under this rule, a registration is entitled to no presumption of validity in court. Secondly, as we observed above in the institutional capacities section, it is possible that Congress can allow authorship of AI. However, it is also important how Congress will make this happen. Congress can add provisions on how the work generated with the help of AI be handled. This amendment can work correspondingly with the human authorship requirement. As we mentioned in the theories section, the user of the AI can be considered as the owner of the work. Congress can also allow it with a subtle change in the language of the statutory text. It can allow authorship to AI. However, granting AI authorship directly might also cause more issues. For instance, after granting authorship to AI, the copyright of AI’s work will lead the questions about the legal personhood of AI. On the other hand, using the doctrines of work-made-for-hire or the theory that the user of the AI is the owner of the work, won’t lead to these questions in the first place. However, eventually the legal personhood or employment capacity of AI will be a crucial discussion because of the autonomous work. For instance, there are AI systems that constantly looking for treatment for diseases and it is still a question of who would be the inventor of these works.[64] The main issue is there is almost no human involvement in the process. The AI has a dataset, and it constantly tries the possibilities to find a cure. Considering the fact that a specific type of AI was made for this research, the developer of the AI can be the owner of the work. Does this mean that a cancer treatment might be patented by an engineer? This issue can also be solved by the work-made-for-hire doctrine by considering private AI as an employee.

CONCLUSION

As we observed in previous sections, the meaning of the words can change over time. What we mean by author is not the literal human being but more of a functional kind. This functional author can create a work that is meaningful and original. Therefore, preventing AI authorship is not helpful in today’s world and is just an obstacle in the way of the goal of the intellectual property clause. If Office is not sure how or to what extent to adopt the AI, rejecting the work altogether is not the solution. As we mentioned in the possible solutions section, the Office can give a temporary position to work while authorities focus on the amendment. The world itself is improving with AI and in a few years it will be at a point where we will rely on AI in every aspect of our lives. Artworks like cinema and music will be dominated by the works of AI. Not allowing anything to these works will just be refusing to adapt to the future. Europe is trying to take steps in this journey, therefore, refusing such adaption will only harm the great authors of the U.S.

[1] B.J. Copeland, Alan Turing and the beginning of AI, Brittanica, https://www.britannica.com/technology/artificial-intelligence/Alan-Turing-and-the-beginning-of-AI

[2] Homage to John McCarthy the father of Artificial Intelligence (AI), teneo.ai, https://www.teneo.ai/blog/homage-to-john-mccarthy-the-father-of-artificial-intelligence-ai

[3] Lawrence B. Solum, Themes from Fallon on Constitutional Theory, 18 Geo. J.L. & Pub. Pol'y 287 (2020)

[4] Author, Merriam-Webster.com, https://www.merriam-webster.com/dictionary/author

[5] Author, Oxford English Dictionary, Historical Thesaurus, https://www.oed.com/search/dictionary/?scope=HistoricalThesaurus&q=author

[6] Inventor, Oxford English Dictionary, Historical Thesaurus, https://www.oed.com/search/dictionary/?scope=HistoricalThesaurus&q=inventor

[7] Eliot T. Tracz, Doctrinal Evolution and the Living Constitution, 42 U. Dayton L. Rev. 257 (2017)

[8] Saqib Shah, ChatGPT 5 release date: what we know about OpenAI’s next chatbot, The Standard, (26 March 2024), https://www.standard.co.uk/news/tech/chatgpt-5-release-date-details-openai-chatbot-b1130369.html

[9] Author, Oxford English Dictionary, https://www.oed.com/search/dictionary/?scope=Entries&q=author

[10] U.S. Copyright Act of 1790

[11] Copyright Office, (2023), Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence, Shira Perlmutter

[12] Id.

[13] Id.

[14] Ryan Abbott & Elizabeth Rothman, Disrupting Creativity: Copyright Law in the Age of Generative Artificial Intelligence, 75 Fla. L. Rev. 1141 (2023)

[15] Shanti Escalante-De Mattei, Stephen Thaler’s Quest to Get His “Autonomous” AI Legally Recognized Could Upend Copyright Law Forever, ART IN AMERICA (January 8, 2024 10:00 am), https://www.artnews.com/art-in-america/features/stephen-thaler-quest-ai-legally-recognized-upend-copyright-law-1234692243/

[16] Id

[17] Thaler v. Vidal, 43 F.4th 1207 (Fed. Cir. 2022), cert. denied, 143 S. Ct. 1783, 215 L. Ed. 2d 671 (2023)

[18] Ryan Abbott & Elizabeth Rothman, Disrupting Creativity: Copyright Law in the Age of Generative Artificial Intelligence, 75 Fla. L. Rev. 1141 (2023)

[19]
Sam Eichner & Aya Hatori, A New Dawn for Copyright in AI-Generated Works?, PILLSBURY: INTERNET + SOCIAL MEDIA (Mar. 9, 2023), https://www.internetandtechnologylaw.com/zarya-copyright-ai-generated-works/ [https://perma.cc/7GCQ-78CW].

[20] Thaler v. Perlmutter, No. CV 22-1564 (BAH), 2023 WL 5333236 (D.D.C. Aug. 18, 2023)

[21] Naruto v. Slater, 888 F.3d 418 (9th Cir. 2018)

[22] Id.

[23] Id.

[24] Id.

[25] Id.

[26] Gia Jung, Do Androids Dream of Copyright?: Examining Ai Copyright Ownership, 35 Berkeley Tech. L.J. 1151 (2020)

[27] Burrow-Giles Lithographic Co. v. Sarony, 111 U.S. 53, 4 S. Ct. 279, 28 L. Ed. 349 (1884)

[28] Id.

[29] COPYRIGHT LAW, IPPRAC MA-CLE 7-1

[30] Lee, Edward, The Code Red for Copyright Law (March 21, 2024). Florida Law Review, Vol. 76, 2024 Forum (Forthcoming), Available at SSRN: https://ssrn.com/abstract=4767791

[31] Id.

[32] Ryan Abbott & Elizabeth Rothman, Disrupting Creativity: Copyright Law in the Age of Generative Artificial Intelligence, 75 Fla. L. Rev. 1141 (2023)

[33] Id.

[34] Abbott, Ryan Benjamin and Rothman, Elizabeth, Disrupting Creativity: Copyright Law in the Age of Generative Artificial Intelligence (August 8, 2022). Fla. L. Rev. Vol. 75, Issue 6, 2023, Available at SSRN: https://ssrn.com/abstract=4185327 or http://dx.doi.org/10.2139/ssrn.4185327

[35] Id.

[36] Id.

[37] Id.

[38] 53 Cumb. L. Rev. 459

[39] Id.

[40] Id.

[41] Id.

[42] Id.

[43] Id.

[44] Padmanabhan, Arjun and Wadsworth, Tanner, A Common Law Theory of Ownership for AI-Created Properties (April 6, 2023). 104 Journal of the Patent and Trademark Office Society 155, Available at SSRN: https://ssrn.com/abstract=4411194

[45] Id.

[46] Id.

[47] Id.

[48] Id.

[49] Id. at 34

[50] Id. at 40

[51] Id. at 40

[52] Id at 34

[53] Id.

[54] Id.

[55] Id.

[56] U.S. Const. art. I, § 8, cl. 8

[57] Edward C. Walterscheid, To Promote the Progress of Science and Useful Arts: The Background and Origin of the Intellectual Property Clause of the United States Constitution, 2 J. Intell. Prop. L. 1 (1994)

[58] Id. At 14

[59] Solum, Lawrence B., Original Public Meaning (April 20, 2024). Michigan State Law Review, Vol. 2023, No. 807, 2023, Available at SSRN: https://ssrn.com/abstract=4801641

[60] Id.

[61] Id.

[62] Id. at 27

[63] Id at 27

[64] Id. at 40