Introduction
IP rights protect the works of creators with economic and moral rights. Copyright legislation protects a literary, dramatic, or artistic work by giving assurance that creators will enjoy protection and certain exceptions. Traditional thought holds that copyrights belong to humans; now with AI, this raises a myriad of questions concerning who owns copyrights.
The core issue that the author seeks to analyse is: Whether legal personhood can be extended to Artificial Intelligence, thus paving a way for attribution of authorship and grant of copyright to them?
The blog is subdivided into two parts. The first part focuses on the jurisprudential analysis of AI-generated works and their personhood, determining whether AI is eligible for copyright attributes. The second part reviews legislative and judicial responses in India and the United States. India’s Copyright Act, 1957, defines an author as a human, which creates challenges for AI-generated works and equally the U.S. Copyright Act, 1976.
In general, the paper develops an analysis of the challenges as well as opportunities that AI brings up for copyright law and contributes to the debate about AI’s future role in intellectual property rights.
1. AI generated works and the question of copyright
1.1 An Overview of traditional Jurisprudence on Personhood, Authorship and Copyright
The legal personality theory asserts the jurisprudential argument for the presence of rights for entities non-human, including copyright and other sorts of intellectual property rights. The personhood theory, attributed to John Locke, Immanuel Kant, and Friedrich Hegel, argues that, for a man to have proper self-development, he ought to have control over external sources. Locke argued that “every man has a property of his own person,” extending to the labour and creations of the individual. Personhood theory therefore posits that authors are intrinsically connected with their works, and dignity for their sense of self demands control over these creations.
Copyright laws in different countries are premised on the theory that rights are bestowed upon natural persons who exemplify originality and labour in their works, in a manner considering those creations as part of one’s personality. Personhood theory is criticised by many. Ralph Waldo Emerson said that “originals are not original”; creative works are derived from or borrowed from other works in several cases. Creativity is inherently accumulative, reflecting lived experiences, inspirations, and adaptations, and thus one contests the uniqueness that personhood theory emphasizes. But the personhood theory remains one of the most basic foundations of copyright law. The rapid development of technology, more especially AI, now ahead of the growth of jurisprudence brings new questions.
1.2 Attribution of Legal Personhood to AI – A Way towards AI Authorship in Copyrights
As Dyschkant suggests, the concept of legal personhood should be decoupled from the notion of humanity. Legally, a person is defined as a subject of rights and duties, and this definition extends beyond humans to include non-human entities like corporations, which are granted legal personality. However, AI systems have not yet been recognized as legal persons, which complicates the regulation of rights and liabilities associated with their use. Recognizing AI as a legal person could streamline rights management and liability issues, thereby enhancing the regulatory framework. The above concept of AI personhood is understood by:
(i) Theories of Juristic Person and Persona Ficta
The Fiction Theory or Realist Theory explains legal status of non-human entities. Under Fiction Theory, the created corporations are fictitious and merely treated as artificial entities by the state. They are attributed with will in order to become a person. The Realist Theory suggests that corporations exist in a pre-legal as well as post-legal context within which they pre-date law; what the law deems as possibly so. Both theories argue to entitle rights and impose duties on non-human entities. Based on these theories, AI can be given a capacity of legal personhood without will-as in the Fiction Theory-or as juristic persons, with human will being attributed for only some purposes.
(ii) Right-Duty Analysis
A legal person is any entity capable of exercising rights and duties. Salmond defines a person as an entity recognised by law for such capabilities. John Chipman Gray extends this view, inasmuch as he holds that entities with either rights or duties qualify as persons. For example, he says that “a right may require a will, but a duty does not require one”. Differently than other non-human entities such as idols or corporations, AI can create creative works by itself. However, human intervention in AI’s programming and data input means that legal personhood can be recognised even when human involvement is minimal. This approach would allow AI-generated works, especially in the case of ‘weak AI’, to receive copyright protection akin to property rights for corporations and partnerships.
2. The Existing Legislative Framework and Judicial Responses to Copyright Challenges in India and the US
2.1 Copyright Laws in India and stance on AI Authorship
The first requirement for copyright to subsist is originality, although the term is not defined under the Copyright Act, 1957. The Supreme Court of India, in a landmark case, adopted the middle ground between ‘sweat of the brow doctrine’ and ‘modicum of creativity doctrine. Section 2(d)(vi) of the Copyright Act, 1957 defines an author in relation to computer-generated work as “the person who causes the work to be created,” and therefore authorship can only be attributed to a person. In Rupendra Kashyap v. Jiwan Publishing House (1996), the Delhi High Court held that authorship is limited to natural persons, though ownership can vest in juristic persons if human involvement is demonstrated.
The Indian Copyright Office briefly bent its head to own up to AI as an author. As recently as 2020, Robust Artificially Intelligent Graphics and Art Visualiser (“RAGHAV”), a piece of art created by IP lawyer Ankit Sahni, gained the right to co-authorship for the first time in AI-generated art. Sahni had initially claimed sole authorship by the AI, which was rejected for the reasons that there was an absence of relevant provisions in the Copyright Act, 1957. However, the Copyright Office sought cancellation of the registration within one year, and that is also being pursued in the United States.
The 161st Parliamentary Standing Committee suggests protecting works generated by AI, either independently or with human aid, by amending existing laws rather than introducing new legislation. This approach sidesteps directly defining AI as an ‘author’, instead focusing on ensuring legal recognition for AI-generated outputs. By recommending adaptable amendments, the committee acknowledges the evolving nature of authorship in the AI era without prematurely resolving philosophical debates. India’s stance reflects a balanced approach, promoting innovation while ensuring intellectual property laws remain robust and flexible to address technological advancements.
2.2 Copyright Laws in the United States and Stance on AI Authorship
Under § 102 of the US Copyright Law in Title 17, the work must be original for copyright to survive. In the Compendium, under § 306, it is declared that the US Copyright Office will only register original works made by a human being. US laws notably require human authorship to entitle works with copyright; consequently, AI-generated work cannot be registered under copyright. In the Creative Machine case, a United States Court of Columbia held that the Copyright Office was right in its decision to deny the registration of visual art created autonomously by the AI since it lacked human authorship. Similarly, in Naruto v. Slater, the US Court of Appeals for the 9th Circuit ruled that a monkey selfie cannot be copyrighted since there is no authorship of a human. In an appeal case involving an invention created by DABUS- an AI- the court confirmed that AI cannot be an inventor under the US patent law because inventorship can only be made of a natural person.
In addition to the many suits that have been filed against Microsoft, another class action suit has been filed by the Authors Guild against OpenAI, alleging to have used works covered with copyrights in training ChatGPT.
In the case of graphic novelist Kris Kashtanova, who collaborated with an AI in creating a graphic novel named “Zarya of the Dawn,” the US Copyright Office initially granted her copyright but then revoked it for each of the images separately once they found the AI origin. Kashtanova had retained copyright over the original story and the arrangement of images.
The US Copyright Office announced an effort to consider questions of law and policy with respect to copyright in AI-generated works. Among these questions are not only issues about use of copyrighted material in training AI but also surrounding, more generally, questions regarding use for AI.
Hence, US copyright law requires human authorship, denying protection for fully AI-generated works while recognizing human contributions in collaborations. Ongoing lawsuits and policy reviews signal the need to adapt intellectual property laws to the challenges posed by AI.
Conclusion and Recommendations
Copyright law is founded on recognizing human creativity as an extension of individuality, which AI inherently lacks. AI operates through pre-programmed algorithms and data analysis, without independent creativity or intent. Granting it authorship would dilute the essence of creativity central to copyright protections. Additionally, recognizing AI as an author raises significant accountability concerns. Unlike humans, AI cannot bear responsibilities or liabilities, creating ambiguity about who should respond in cases of copyright disputes or misuse, potentially resulting in regulatory loopholes. Economic and moral rights also come into play; AI systems are tools created, owned, and controlled by humans, and the benefits of AI-generated works should belong to the individuals or entities who develop and maintain these systems.
Furthermore, moral rights, which are personal to human authors, cannot be attributed to AI since it lacks emotions, beliefs, or ethical considerations. Restricting AI authorship does not hinder innovation but instead encourages collaborative efforts where humans retain authorship for their contributions, ensuring fair use while preserving the integrity of copyright law. Precedents set by courts worldwide, including in landmark cases such as Creativity Machine, Naruto v. Slater, and DABUS, have consistently rejected the extension of authorship to non-human entities. Upholding this stance maintains legal consistency and avoids destabilizing intellectual property frameworks. Restricting AI authorship balances innovation with the protection of intellectual property, ensuring the law adapts responsibly to technological advancements while preserving its foundational principles of originality, human creativity, and accountability.
The copyright of AI-generated works is a complex issue, as many jurisdictions struggle with the role of non-human creators in intellectual property. Legislative reform could address these challenges by introducing a ‘synthetic authorship’ framework that grants copyright to the individuals or entities controlling the AI’s creative process, such as developers or users. Alternatively, works created purely by AI without significant human intervention could default to the public domain, ensuring free access and preventing monopolization of creative resources. A potential criterion for assessing human involvement could involve a modified ‘Turing Test’ for creativity, where the work is evaluated to determine if it exhibits creative input indistinguishable from human effort. Revisions to fair use or fair dealing laws may also be necessary to balance the interests of original human authors and users of AI-generated works. Clear guidelines on liability and moral rights for AI developers or dataset creators, coupled with a compulsory licensing model for AI tools, could foster transparency, accountability, and equitable use.
Different jurisdictions offer insights that could guide such reforms. The United States Copyright Office does not grant copyright protection to works created solely by non-human authors. In Thaler v. Perlmutter, the court confirmed that human authorship is a fundamental requirement for copyright. and the UK provides a limited form of copyright protection under Section 9(3) of the Copyright, Designs, and Patents Act, 1988, for “computer-generated works” with no human author. The ‘author’ is defined as the person who made the arrangements necessary for the work’s creation.
Sneha Agarwal, 3rd Year, National Law Institute University, Bhopal
