Artificial Intelligence And The Future Of Property Law - Authorship, Ownership And Legal Recognition
Author: LawSphereX
Category: Technology, Cyber & Privacy
Published:
Updated:
ARTIFICIAL INTELLIGENCE AND THE FUTURE OF
PROPERTY LAW: AUTHORSHIP, OWNERSHIP AND
LEGAL RECOGNITION
Raewa Bharat Parab, O.P. Jindal Global University
ABSTRACT
Artificial intelligence (AI) in the present age has started to disrupt the traditional legal meaning of authorship, ownership, and property due to its rapid development. The regimes of intellectual property with their longstanding history based on human agency and creativity are now challenged by the unprecedented need to accommodate the machine-generated outputs that have economic and originality value and commercial usefulness. This paper discusses whether AI-generated creations can qualify as property under natural law theories and legal doctrines that are already in existence. It also discusses the consequences of the silence of law, specifically the ‘Transfer of Property Act’, 1882,1 the ‘Copyright Act’, 19572 and the ‘Patents act 1970’,3 and examines the effects of these gaps on making AIgenerated works legally recognized.
The paper argues that although AI cannot be identified as a legal subject matter, AI-generated works ought to be treated as property, as long as sufficient human contribution can be established. AI cannot be given the sole status of an owner or author and therefore, the author argues for joinownership or authorship rather than complete neglect. Lack of legislative clarity in India leaves everyone confused, yet the statutory material provides flexibility in interpretation. The legal principle of recognizing AI-generated works as property is not only the necessary legal step but the logical progression of property law according to the technological change.
1 Transfer of Property Act 1882.
2 Copyright Act 1957.
3 Patents Act 1970
CHAPTER 1: INTRODUCTION
Artificial Intelligence has propelled from the realm of technology into the legal arena, thereby becoming a matter of judicial scrutiny. Ankit Sahni (RAGHAV) v UOI4 is a landmark case which deals with ambiguity on ownership of AI generated creations and AI as a co-author. The Delhi High Court and legal thinkers are grappling with a deep-seated conflict within Intellectual Property Law on whether the author should be a natural person or if AI assisted work, with prompts given by humans should be given protection.
1.1 PROPERTY AND EVOLUTION OF TECHNOLOGY
Property is that despotic dominion over the exclusion of others. For centuries, property has been defined as a legally sanctioned bundle of rights with the most notable ones being the right to exclude, right to use, right to transfer and the right to derive economic benefit. Scholars like Merrill argue that the “Right to Exclude” is imperative without which property comes under public domain.5 As per the Indian law, property law distinguishes between moveable and immoveable property under the ‘Transfer of Property Act’, 18826 and recognizes intangible property through specialized statutes such as the ‘Copyrights Act’, 19576 and the ‘Patents Act’,
1970.8 However, the definition of ‘property’ remains the same across all statutes.
Rapid technological advancement has led to an increasing rise in deployment of AI which poses an unprecedented challenge in the field of law. Unlike earlier tools and systems, modern generative AI is now capable of not only assisting human created work but also has the ability to independently produce literature, art, music, making it difficult to distinguish between machine and human creativity. They exhibit characteristics of what is considered as ‘property’. Yet the law remains silent on whether these AI generated creations constitute property and moreover, on who owns them since IP law, historically, has been centered on human creativity only. This silence creates a doctrinal vacuum at the intersection of law and technology causing regulatory uncertainty which leads us to the question “What constitutes property in the age of AI?”.
4 Ankit Sahni v Union of India
5 Thomas W Merrill, 'Accession and Original Ownership' (2009) 1 Journal of Legal Analysis 459.
6 Transfer of Property Act 1882 (n 1).
6 Copyright Act 1957 (n 2).
8 Patents Act 1970.
1.2 RESEARCH QUESTIONS
1. Will the theoretical requirement to be deemed as property be fulfilled by AI generated
creations?
2. What are the obstacles posed by the existing legal frameworks and can they be
overcome?
3. Who owns AI generated creations and can AI be considered an author?
4. Is there any country that recognizes AI generated creations as property?
1.3 HYPOTHESIS
The central hypothesis of the paper is that AI generated creations should be considered property through legislation with sufficient human contribution due to the growing advancement and dependence on technology and argues for joint ownership.
1.4 RESEARCH METHODOLOGY
The paper discusses these ambiguities through a doctrinal, theoretical and comparative analysis. The paper analyses legal statutes such as the ‘Transfer of Property Act’, 1882, the ‘Copyright Act’, 1957 and the ‘Patents Act’ 1970 and the provisions of the foreign laws, namely, the ‘Copyright, Designs and Patents Act’, 1988 of the UK and new emerging legal regulations in jurisdictions as China and Ukraine. Judicial decisions and case laws have been critically engaged with in order to grasp how courts comprehend the concept of ownership, authorship and originality in the context of AI generated creations. Besides statutory interpretation, the paper also pays significant heed to the theoretical frameworks to assess the extent to which AI-generated creations address the principles of the law of property.
A comparative methodology has been utilized to find similarities, differences and reforms in jurisdictions to allow a more accurate view of how various legal systems are reacting to AIbased challenges. Secondary sources, such as scholarly articles, law review commentaries, blogs, books and journals have also been used to substantiate doctrinal arguments and offer a critical view. The study is analytical in nature, and it aims at interpreting and evaluating legal principles. The multi-layered methodological approach enables the paper to close the gap between theory and practice and provide an in-depth and prospective analysis of the
AIgenerated creations in the context of the changing environment in property law.
1.5 LITERATURE REVIEW
1. ‘Transfer of Property Act’, 1882- Provides the framework of transfer of property in
India, mostly in regard to tangible and immovable property, with a broader acceptance of transferable interests. It does not specifically refer to intangible or digital properties, and it leaves a loophole to define AI-generated works as property.
2. ‘Copyrights Acts’, 1957- Protects original literary, artistic, and musical works, and
authorship is based on the human creativity. Here it is being critically analysed to see if AI generated creations can constitute property.
3. ‘Patents Act’, 1970- Deals with the concept of inventions and inventors. It denies the
inventorship to AI systems, which poses a structural challenge to accepting AIgenerated inventions as property.
4. ‘Copyright, Designs and Patents Act’, 1988- A UK legislation used for comparative
analysis with Indian legislation with it primary focus on section 9(3) to understand if UK legally recognises AI generated creations as property.
5. William Blackstone- This text talks about the traditional meaning of property. His
model is the basis of contemporary property law but presupposes the ownership of humans, which becomes problematic when it comes to the work of AI.
6. “Accession and Original Ownership” by Thomas w. Merrill- He discusses the
characteristics of property, the ways ownership is created by change and domination of resources. His theory advocates the expansion of the property rights to AI outputs in situations when the human actors control the creative process.
7. “John Locke’s Labour Theory” by Aravind Prasanna- Elaborates that a property
emerges when people combine their labour and usual resources. This offers grounds to consider AI-generated work as property in terms of the labour of developers and users.
8. “A Lockean Theory of Intellectual Property” by Adam D. Moore- Engages with how
labour when AI is involved is multi layered i.e. it makes use of the developer as well as
the prompter’s skill and labour.
9. “An Introduction to the Principles of Morals and Legislation” by Jeremy Bentham-
Uses the utilitarian theory along with principles of economic incentives where property is considered to be a means of maximizing social welfare. IP protection of AI-generated works as a tool to stimulate innovation and economic development.
10. 'The Philosophy of Intellectual Property' by Justin Hughes- Combines labour, utilitarian
and personality theories to explain the rationale of IP rights. Offers a detailed system of approaches to the study of AI-generated works in the current property doctrines.
11. Elements of the Philosophy of Right by GWF Hegel- Formulates the theory of
personhood, which associates property with self-expression and self-identity. Although AI does not have a personality, creative input of humans is encouraged in the theory AI-generated content.
12. 'Analysis of Doctrines: "Sweat of the Brow" & "Modicum of Creativity" vis-à-vis
Originality in Copyright Law' by Shuchi Mehta- Examines criteria of originality of copyright law. Shows that protection can be justified by even the slightest human work, which is an argument in Favor of the recognition of AI-guided outputs.
13. 'Artificial Intelligence and Intellectual Property' by WIPO- Looks at the models of
ownership of computer-generated works with a focus on human involvement. Proposes loose allocation models, which affect the discussion of AI ownership.
14. 'Allocating Ownership Rights in Computer-Generated Works' by Pamela Samuelson-
15. 'IP in a World Without Scarcity' by Mark A Lemly- Claims that the abundance of digital
content is a problem to traditional IP arguments. The necessity to raise the issue of overprophetization and at the same time the need to protect the AI in a balanced manner.
16. 'Artificial Intelligence & Copyright: Section 9(3) or Authorship without an Author' by
Toby Bond and Sarah Blair- Critically evaluates the UK attitude to AI authorship. Emphasizes the ambiguity and restriction of attributing authorship with the help of the necessary arrangements.
17. 'Copyright Protection for AI-Generated Outputs: The Experience from China' by Yong
Wan and Hongxuyang Lu- Researches the case-based approach in China, which focuses on human intellectual contribution. Shows that the judiciary has been flexible in accepting AI-assisted works as being copyrightable.
18. “Law of Ukraine on Copyright and Related Rights”- Brings forward the provisions
regarding computer-generated works and new digital assets. Recommends a move to sui generis protection without abating human authorship conditions.
19. Ankit Sahni v Union of Indi’- Deals with ambiguity on who owns AI generated creations
and if AI should be given the status of a co-author.
20. Li Yunkai v Liu Yuanchun- The court held that human intellectual effort is used to create
AI generated works and thus it should be copyrighted.
21. Eastern Book Company v. D.B. Modak- Supports the definition in the copyright act
where the author of a work should be human and the Supreme Court substantiated it using the “modicum of creativity” criterion by stressing on skill and judgment as the key features of originality.
22. Walter v Lane- Applied the doctrine of “sweat of the brow” and granted copyright to
reporters for verbatim reports of speeches due to their skill, labor and expense which made them “authors” of the published reports.
23. Burlington Home Shopping Pvt Ltd v Rajnish Chibber- Case in the Indian Jurisdiction
strictly adhering to the “Sweat of the Brow” doctrine.
24. Thaler v Comptroller-General of Patents, Designs and Trade Marks- Demonstrated the
strict interpretation of the Patents Act where non-human inventors were not given recognition and how this poses a challenge in the current times.
CHAPTER 2: THEORETICAL FOUNDATIONS OF PROPERTY
The categorization of AI-created works as property involves interactions with classical theories of property and intellectual property which provide a strong framework to argue for AI generated creations to be constituted as property and that AI could be recognized as co-owners
or co-authors. These theories help bridge the gap between traditional legal doctrines and the modern reality of autonomous machine creation. Although all these frameworks were constructed within human creativity, they can be redefined to fit the technologically mediated creation.
2.1 JOHN LOCKE’S THEORY
John Locke’s theory, ‘the labor theory of ownership’, is one such theory which can be used to support intellectual property, which argues that people acquire property rights by mixing their labor with unowned resources from the “commons”. 7 The primary assumption made by Locke is that since an individual’s body is their own property and that they own its labor, they are entitled to the final product of that intellectual labor8. Copyright in ‘section 13 of the Copyright Act’, 1957, is recognized in “original literary, dramatic, musical and artistic works”9 where Indian jurisprudence interprets originally as something that requires skill, labor and judgement. However, in AI’s context, there are several layers to this labor, implying that the labor is multilayered.
The first layer of labor comprises of the work of developers that design and direct the algorithms and filter huge datasets to be used during the training. Second, there are the “labor” of the human prompter that offers direction to generate a given output. 10 This demonstrates that although the output is machine based, human intellect is used to generate it. However, “Lockean theory of appropriation was found to be influential, as propounded by HUGO Grotius since the former argued that the labor that is exerted upon the resources creates greater morally binding restrictions than any social contractual obligations. ”11 This version of the theory provides a stronger foundation for providing rights to creators of intellectual property.
“According to him, the first person who employs his or her labor to the resources available has the sole right to appropriate it without anyone else’s consent. ”12 Locke justifies this theory of appropriation by categorizing it into five points, with the first one being that “when a person owns their labor and mixes it with things which are not owned with their labor, 7 Aravind Prasanna, 'John Locke's Labour Theory: A Justification of IPRs' (Legal Services India, 2020) accessed 3 March 2026.
8 ibid 10
9 Copyright Act 1957 (n 2) s 13.
10 Adam D Moore, 'A Lockean Theory of Intellectual Property' (1997) 21 Hamline Law Review 65.
11 Aravind Prasanna, 'John Locke's Labour Theory: A Justification of IPRs' (Legal Services India, 2020) accessed 3 March 2026.
12 ibid 14
he or she becomes the self-owner of the property.”13
2.2 UTILITARIAN THEORY
Utilitarian justification provides a different perspective in understanding why AI-generated creations should be regarded as property by our legal frameworks. The fundamental argument behind the utilitarian theory is that property is not a natural right, but rather a practical device that aims at maximizing the aggregate social welfare by encouraging innovation and creativity, limiting monopolies to creators as incentives to invent. 14 The intellectual property regimes and one of which is the Copyright Act, 1957 are structured in such a manner that they provide exclusive rights to the creators in an attempt to encourage the creation and spread of socially desirable works. In the scenario of artificial intelligence, this argument gains even greater importance.
If AI outputs remain in the public domain by default, developers may have insufficient incentive to invent while competitors have an excessive incentive to copy. 15 Closely connected with this is the principle of economic incentives, a concept that serves as the practical application of utilitarian logic. The digital economy, day by day, is becoming increasingly based on the AI-generated outputs in industries such as the sphere of entertainment and journalism, software development, and medication. These are not just theoretical constructs, but commercially exploited products that have quantifiable market value. The design, education, and execution of generative AI applications are costly in terms of capital, time, skills, and infrastructure. A lack of protection for AI generated work could lead to lost opportunity and a no-protection rule of secrecy.
16 Moreover, due to the lack of ownership rights, licensing, commercialization, and enforcement are difficult, which causes uncertainty of the law and inefficiency of the market. Utilitarianism wise, it is important to acknowledge that AI-generated creations are property to ensure that the ecosystem in which technology is developed is sustainable to achieve greater societal good and for the general goal of creating economic growth and innovation through creating legal certainty.
13 ibid 14
14 Jeremy Bentham, An Introduction to the Principles of Morals and Legislation (first published 1789, Clarendon Press 1996). 15 Edwin C Hettinger, 'Justifying Intellectual Property' (1989) 18(1) Philosophy & Public Affairs 31. 16 Adam Moore and Ken Himma, 'Intellectual Property' in Edward N Zalta (ed), Stanford Encyclopedia of Philosophy (Spring 2022 Edition, Metaphysics Research Lab, Stanford University, 2022) accessed 12 March 2026. 19 GWF Hegel, Elements of the Philosophy of Right (first published 1820, Allen W Wood ed, HB Nisbet tr, Cambridge University Press 1991).
2.3 PERSONHOOD THEORY
Personhood theory, which is inspired by Hegel and Kant, asserts that property is a continuation of individual personality and is a key to self-actualization19 since individuals identify themselves through the combination of their talents, feelings and character traits with external objects. Within this perspective, the intellectual objects that are within the control of an author enable the will of an author to be manifested in the external world. Although AI does not have a personality or a will, the pieces of work it produces tends to be highly suggestive of the creative selections of the human prompter. In this regard, the AI is an advanced mechanism or a protrusion of the will of the human creator. The artist employs the AI to safeguard their personal life or support the lifelong work.
When we accept that the authors are taking some risks when they place themselves on display by their creations, the intellectual property rights will provide them with some reasonable degree of control over that risk. 17 This ethical tether between the creator and the work relates to the premise that the human operator must be accepted as a co-author or the work of the AI should be regarded as intellectual property. The case, Li v. Liu in China,21 was an important judgement. The court ruled that more than 150 prompts were a personalized expression of the author with a personal touch to the image generated by AI. When the law recognizes that the authors take risks by placing themselves on display with their works, the intellectual property right extends to them providing a requisite amount of regulation over such risk.
2.4 DOCTRINE OF SWEAT OF THE BROW
Although AI is directly not stated in Indian law, the doctrine of “sweat of the brow” also known as the “Industrious Collection”,18 posits that an author’s effort, investment and diligence is what helps them acquire property rights implying that creativity or originally of an author may be insignificant. This doctrine has historical roots in the English law, often linked to the 1990’s case of Walter v Lane23 which granted copyright to reporters for verbatim reports of speeches due to their skill, labor and expense which made them “authors” of the published reports. When applied to AI generated creations, this framework offers compelling arguments for constituting
17 Justin Hughes, 'The Philosophy of Intellectual Property' (1988) 77 Georgetown Law Journal 287.
21 Li Yunkai v Liu Yuanchun (2023) Jing 0491 Min Chu No 11279 (Beijing Internet Court, 2023).
18 Shuchi Mehta, 'Analysis of Doctrines: "Sweat of the Brow" & "Modicum of Creativity" vis-à-vis
Originality in Copyright Law' (IndiaLaw, 8 January 2015) accessed 4 March 2026. 23 Walter v Lane [1900] AC 539 (HL).
such outputs as property. In this context, the computational effort to run through millions of permutations and combinations to discover a desired output may be a modernized version of the industrious collection. In the past, Indian courts were very vigorous in adhering to the sweat of the brow approach in such cases as Burlington Home Shopping v. Rajnish Chibber,19 where it was held that even though compilation of database creates originality that is next to negligent, the law protects it and since it involves diligence and an enormous amount of time and effort to put it together and no man has the right to appropriate himself for another person’s brain, skill or labor.
The legal system should deem AI outputs a property, given this industrious labor, which would encourage even further development of transformative technologies that would otherwise go under-produced by the threat of too much incentive to copy or literary piracy. This doctrine in finality provides the multi-layered ownership structure, in which efforts of both developers and users are considered to be the source of the property in order to ensure that the creations are not in a doctrinal vacuum whereby they are left without owners and unlawful. The doctrine of the sweat of the brow offers a relaxed and practical criterion of recognition of the value generated through human-machine cooperation in a silence-based legal system.
CHAPTER 3: LIMITS OF THE EXISTING FRAMEWORK
3.1 TRANSFER OF PROPERTY ACT, 1882
The Transfer of Property Act (TPA), 1882 is the primary instrument used for transferring of rights between the parties which include humans and corporations but is silent on the status of non-human beings. Section 520 of the Act describes a transfer of property as an act according to which a living person transfers property to another one. Though the Act does not comprehensively provide a meaning of property, judicial interpretation has always taken a comprehensive approach to the definition of property, and it includes any interest that can be possessed, enjoyed, or conveyed. The TPA is essentially organized in terms of material property especially the immovable property like land and buildings.
Although it does not formally rule out intangible property, the creations of AI under the bifurcation of the TPA would have to be considered intangible movable property because they are creations of the mind and cannot be classified under the definition of immovable property in the statute which is comparable to the intellectual property. They are not physical but have economic value and
19 Burlington Home Shopping Pvt Ltd v Rajnish Chibber 1995 PTC (15) 278 (Delhi HC).
20 Transfer of Property Act 1882 (n 1) s 5.
can be licensed. But because AI does not have a legal personality, it cannot technically contract and assign the proprietary interests to a human owner under the TPA. This leads to a legal contradiction where a valuable piece of property exists without any owner due to the existing legislations, and this is conceptually unsustainable in property law. One of the most significant issues that legal thinkers have provided is the classification of these works. Personal property historically falls into two categories: things in possession (physical objects) and things in action (intangible rights, which may be enforced in court).
Critics state that the creations of AI do not fall easily into either category: they are neither things in possession since they are not physically established, nor things in action since they can exist without a debt or a legally binding obligation to a particular counterparty. 21 Several scholars, and other organizations, including the Financial Markets Law Committee (FMLC), propose the introduction of a Third Category of personal property to address this problem. 22 In this category, the digital assets would be recognized as virtual things, which are rivalrous (use by one prevents use by others) and can have exclusive factual control, although the possession does not exist physically.
Although the TPA acknowledges property in general, it, does not directly deal with such types of property and it is up to special laws such as the Copyright Act to deal with them, making such creations a grey area in law.
3.2 COPYRIGHT ACT, 1957
The copyright act 1957 is the main law regarding intangible creative work in India. Under Section 13,23 “the copyright exists in original literary, dramatic, musical and artistic works”. The Act however is founded on the concept of authorship which has been defined in ‘Section 2(d)(iv)’.24 ‘Section 2(d)(vi) of the Copyrights Act’, 1957 states that “the person who causes the work to be created” shall be regarded as the author in case of computer-generated work, implying that human authorship is a pre-requisite for copyright protection. However, the act does not explicitly address AI generated outputs. This provision at first sight does seem to support non-traditional forms of creation. It was, however, passed in a pre-AI era, and does not directly consider autonomous generative systems. The ambiguity of the sentence is in the fact
21 Mark A Lemley, 'IP in a World Without Scarcity' (2015) 90 New York University Law Review 460
accessed 6 March 2026. 22 Financial Markets Law Committee, 'Response: Law Commission Call for Evidence on Digital Assets' (FMLC, July 2021) accessed 6 March 2026.
23 Copyright Act 1957 (n 2) s 13.
24 Copyright Act 1957 (n 2).
that it says that the ‘work is created because of something’. Within the case of AI, several actors can take this role which include the developer who wrote the algorithm or the entity who trained the model or the user who gave prompts. The law fails to give any directions on how to settle such conflicting claims and nor does it explain the extent to which the authorial credit can be confirmed in case of minimal human input like a mere prompt. This dilemma poses a major setback. The property rights cannot be acknowledged in the absence of ownership.
3.3 PATENTS ACT, 1970
The Patents Act, 1970 poses a serious doctrinal challenge to consider AI-generated creations as property, mainly because of its anthropocentric vision of what it means to be an inventor or a proprietor of a creation. 25 ‘Section 2(1)(y)’31 defines a “true and first inventor”26 as a person who is the real deviser of the invention and expressly excludes a first importer of an invention into India but implicitly takes it to assume the existence of a natural person who can be an intellectual deviser. Such a definition places an immediate obstacle to inventions made by AI, because even though the artificial intelligence systems have the ability to autonomously produce new results, they cannot be legally considered inventors under the statutory provision.
This necessity is also supported by the fact that the right to file a patent can be limited to the true and first inventor, an assignee of the true and first inventor, or a legal representative, thus structurally barring non-human entities to the patent system. This restriction has some practical implications. In case an invention is produced completely by AI with no trace of a human input that passes the threshold of inventorship, an invention may end up in a grey area where no valid patent application can be filed leaving it without a property owner. This narrow scope of the legislation is transferred into the administrative and judicial practice. An Indian Patent Office in the case of Thaler v.
Controller General of Patents, Designs and Trade Marks (2023)27 refused to accept a patent application in which a non-human inventor was named (DABUS), stating that the Patents Act did not support non-human inventors. The ruling clarified that the wording of the statute is categorical that it must be an invention of a human being, which puts India in line with other jurisdictions. All these decisions indicate a worldwide hesitation to grant inventorship rights to AI based on the fear of legal personality,
25 World Intellectual Property Organization, 'Artificial Intelligence and Intellectual Property'
(WIPO) accessed 13 March 2026. 31 Patents Act 1970
26 ibid 32
27 Thaler v Comptroller-General of Patents, Designs and Trade Marks [2023] UKSC 49 (20 December 2023).
responsibility and the principles of patent law. In addition, the need of an inventive step, as set out in Section 2(1)(ja),28 which is characterized as a feature by way of technical progress or economic value and is not apparent to a person who is skilled in the art, implicitly presupposes a human standard of skill and creativity. 29 This creates interpretive difficulties in evaluating AI-created inventions because the criterion is not easily adjusted to machine-driven innovation that might be more effective than human innovators. Moreover, the patent system is also intertwined with the concept of disclosure and reward, in which exclusive rights are awarded in the case of disclosed invention to the population. But with AI, where the results can be produced by opaque black box mechanisms, it becomes difficult to assign a creative input and meaningful disclosure.
30 The failure to establish a human inventor does not only interfere with attribution of rights but also the enforcement and transferability of right which are central to establishing inventions as property. In its present state, therefore, the Patents Act, 1970, poses a structural and conceptual impediment to the acceptance of AI-generated works as property, as this would require either a deliberate reinterpretation or a legislative change to reflect the realities of machine-generated innovation.
3.4 INTERPRETATION OF STATUTES
These Act when interpreted strictly and literally would prevent the recognition of the AIgenerated creations as property as its focus would be on fulfilling the act’s requirement of a living person. The copyright act assumes that the author of the work must be a human. The judicial interpretation supports this stand especially in Eastern Book Company v. D. B. Modak31 where the Supreme Court used the “modicum of creativity” criterion by stressing on skill and judgment as the key features of originality. But literal sense does not suit well in resolving the technological advancements that were not considered in the moment when the law was made. Purposive interpretation has been taken numerous times by the Indian courts to make statutes applicable in the changing contexts. Purposive interpretation allows the intentions behind the legislation to be seen and not the words themselves.
Under the intellectual property law, the main goals would be to encourage innovation and creativity, secure economic interests and
28 Patents Act 1970
29 World Intellectual Property Organization, 'Artificial Intelligence and Intellectual Property' (WIPO) accessed 13 March 2026. 30 Enrico Bonadio, Eduardo Alonso and Vansh Tayal, 'The AI Black Box Issue and Patent Disclosure' (City Law Forum, 6 May 2025) accessed 13 March 2026.
31 Eastern Book Company v DB Modak (2008) 1 SCC 1 (SC).
appropriate distribution of rights. By interpreting the provision of ‘Section 2(d)(vi) of the Copyright Act’, 1957, purposively, it is possible to have an expanded interpretation of the meaning of the person who causes the work to be created. This could include the AI system directed by the user or the person that created the generative structure or even a combination of both. Similarly, the general understanding of property as it is in the TPA could potentially be applied to intangible digital property, such as AI-created works, provided they have an economic value and can be transferred. The Indian jurisprudence has long been flexible regarding the adaptation of legal concept to the changing technology. An example is software and digital assets, which have been considered property even though they are not directly stated under more ancient laws.
Fundamentally, courts can be vital through purposive interpretation so that the provisions that exist can be applied to AI-conceived works. By defining such outputs as property, the courts can be in a position to offer an immediate clarification as opposed to allowing legislative reform to prevail. The judicial or administrative practices can be used to explain that authorship under Section 2(d)(vi)32 encompasses persons who exercise a material control or direction to the AI systems. This would give a feasible criterion of ownership. 33 The Parliament can discuss the revision of the Copyright Act to directly discuss the works created by AI. This could include defining AIgenerated work, setting up authorship standards, clarifying ownership rules or finding a middle ground to connect it to the act.
34 The TPA may be read or revised so as to expressly acknowledge the intangible digital property as a transferable property. This would bring property law to the modern technological realities. A joint- ownership or co- authorship model could be suitable in situations where there is more than one contributor. This strategy indicates the decentralized character of AI development and fair distribution of rights. Although the purposive interpretation provides a feasible way, it is not devoid of constraints. Over application of judicial interpretation can create inconsistency and uncertainty. A legislative intervention is more explicit but can become out-of-date in technology. Besides, the extension of property rights to AI-generated work provokes the question of over-propertization and monopolization. The fact that the Indian law is silent on the AI-generated creations does not
32 Copyright Act 1957 (n 2) s 2(d)(vi).
33 Pamela Samuelson, 'Allocating Ownership Rights in Computer-Generated Works' (1985–1986) 47 University of Pittsburgh Law Review 1185. 34 Thushar V Senan, Abey Augustine and Aswathy Krishnan, 'AI, Creativity, and Copyright Law in India: Navigating the Boundaries of Originality and Authorship' (2023) 6(3) International Journal of Law Management & Humanities 2941 accessed 15 March 2026.
indicate non-exclusion but just a restricted scope of statutory foresight. The conceptual flexibility to enable the new forms of property is present in the Transfer of Property Act, 1882 and Copyright Act, 1957, but neither of them is specifically oriented towards the digital age. The challenges that legislation silence creates, including uncertainty, difficulty in enforcement, and inconsistency in the doctrine, are solvable by purposive interpretation, judicial innovation and specific legislative reform.
CHAPTER 4: AUTHORSHIP AND OWNERSHIP OVER AI GENERATED
CREATIONS
Silence by both laws in this context pose some practical and doctrinal issues. The fundamental one being uncertainty in ownership. There is a lack of clear statutory guidance and thus conflicting claims regarding AI-generated outputs. This will not encourage investment and will make commercialization difficult. It might not be easy to trust AI-generated content because businesses might not be ready to lose ownership. Contrary to the tradition, where one can identify and pinpoint an actual human creator of a work, AI-generated artworks are a result of a complex and decentralized procedure that includes various participants. These are usually the designer and trainer of the AI system, the prompter who feeds the system and the machine itself which produces the end product by a series of algorithmic steps.
35 In this part, a critical analysis of each of the possible claimants has been made and the idea of joint ownership or coauthorship has been considered as providing a doctrinally correct and practically feasible solution. The most extreme suggestion could be that AI system ought to be considered an author or a proprietor of the work that it produces itself. The argument can be based on the fact that AI systems are becoming more autonomous and can generate outputs without the direct intervention of humans. However, legal and even philosophical challenges exist that make it impossible to overcome this stance. According to the Indian law, the position of property can only be held by legal entities that are both natural and juridical entities such as human beings or corporations. AI systems are neither legally a person nor can they hold any rights or responsibilities.
Additionally, the fact that AI is an author will interfere with the intellectual property law framework. Authorship involves the right to own, as well as the ability to assign, license and exercise rights, which AI systems cannot do on their own. According to Pamela Samuelson, the intellectual property law has a rationale that is based on human agency and 35 Aravind Prasanna, 'John Locke's Labour Theory: A Justification of IPRs' (Legal Services India, 2020) accessed 3 March 2026.
responsibility. 36 This claim is grounded on the fact that the developer plays a significant role in designing, training, and implementing the AI system. This is an area of heavy intellectual and economic investment, which is in line with Lockean ideas of labor as source of ownership. The nature and extent of outputs that the system can generate is undoubtedly determined by the architecture, datasets and algorithms that are incorporated into the system. This assertion however becomes less powerful when considered in connection to the uncertainty and independence of AI outputs in the real world. When AI systems become operational, they produce content based on various and changing user input which may not be foreseen or controlled by the developer.
Allowing developers to claim such outputs as blanket copyright is likely to run the risk of over-extending the proprietary rights and could lead to monopolistic control over an enormous and indefinite mass of content. 37 Utilitarians would find this suppressive to downstream creativity and limit the wider access to innovation, which is the same goal intellectual property law intends to accomplish. On the other hand, the user or the prompter gives a closer relationship to the output that is generated, especially through the process of initiating and guiding the creative process. The user has some level of control which could be perceived as creative input by framing prompts and making choices on outputs and refining results.
‘Section 2(d)(vi) of the Copyright Act’, 1957, through purposive reading, might bear out the attribution of authorship to such persons who cause the work to be created, particularly where their involvement is indicative of intellectual effort. Prompts in a lot of cases contain the aesthetic judgment, conceptual guidance, or even style preference, thus providing some personal touch to the end product. 38 Such a stance is not without its restrictions, however. Not every user interaction qualifies as meaningful creativity; where the inputs are low or generic, there is a danger that extending authorship will weaken the barrier of originality. Moreover, the user contribution is taken as an exclusive concern, and the enabling nature of the developers is easily overlooked, whose systems make such outputs possible in the first place.
Since it is impossible to assign either developers or users the exclusive ownership, the 36 Pamela Samuelson, 'Allocating Ownership Rights in Computer-Generated Works' (1985–1986) 47 University of Pittsburgh Law Review 1185.
37 Mark A Lemley, 'IP in a World Without Scarcity' (2015) 90 New York University Law Review 460
accessed 20 March 2026.
38 J Jane C Ginsburg, 'The Concept of Authorship in Comparative Copyright Law' (2003) 52 DePaul Law
Review 1063 DePaul University accessed 20 March 2026.
idea of joint ownership or co-authorship may be deemed most appropriate. This model acknowledges the dispersed aspect of creativity in the work of AI-generated and strives to assign rights in a much more equal way across the participating actors. ‘Section 2(z) of the Copyright Act’, 1957, “defines a work of joint authorship as a work in which two or more authors collaborated, and the contribution of one author is not obvious over the other”. Although the given provision was not tailored to AI contexts, its key idea can be extracted. In the example of AI-created works, the technological infrastructure is provided by developers. Creative direction is inputted by users. The interaction is mediated by the AI system. The contributions are separate in nature, but they complement one another. The end product cannot be solely attributed to a particular actor.
A joint ownership model is compatible with Lockean theory through the sources of labor, Utilitarian theory, through the allocation of incentives among donors and Personhood theory, in its ability to maintain the role of human expression. But joint ownership is hard to put into practice. It might be challenging to define how much each party contributes and rights distribution issues could be demonstrated. However, they are not specific to AI, and such difficulties prevail in conventional collaborative productions, and they can be handled by contractual agreement. Practically, the dispute over ownership can be easily solved with the help of the contractual agreements. The developers can provide terms of use that grant out the rights to the users, reserve some rights or come up with shared ownership. These contracts are flexible and able to conform to various uses.
However, the use of independent contracts is not enough. There is a need to have an explicit legal framework in order to bring about uniformity and guard weaker parties. 39 All ownership models have their advantages and disadvantages. Here, the author AI is legally untenable. The developer and the user cannot claim full authorship and the machine cannot be a legal person. This proves that the idea of joint ownership offers a practical solution by conforming to the theoretical foundations of property and adapting to the growing technology. It incorporates the cooperative approach to AI development and maintains the anthropocentric basis of the property law.
CHAPTER 5: COMPARATIVE ANALYSIS
5.1 UNITED KINGDOM
39 Robert Burrell and Allison Coleman, Copyright Exceptions: The Digital Impact (Cambridge University Press 2005).
The United Kingdom takes one of the most direct statutory methods of AI-generated works, namely, by the application of ‘Section 9(3) of the Copyright, Designs and Patents Act 1988 (CDPA)’,40 “which states that in the instance of computer-generated works, the author shall be deemed to be the person who makes the arrangements by which the work is created. ” This clause is a conscious theological break with the conventional human-centric authorship, which brings in a legal fiction to permit copyright to exist even in the face of no direct human creativity. 41 In contrast to systems that are based on interpretive expansion, the UK system tries to actively accommodate the technological change by making sure that outputs of AI should not be left out of protection just because they do not have a natural author. Nevertheless, this legal clarity covers more conceptual ambiguity.
The term “arrangements necessary” is not specified which contributes to the ambiguity as to whether the developer, user or an intermediate party is the author of the arrangements necessary. This ambiguity is becoming increasingly prominent in the presence of generative AI systems, where the interaction between people and machine is extremely diluted. The UK copyright law, based on the European jurisprudence, notably ‘Infopaq International A/S v Danske Dagblades Forening’,42 must show that a work can and does represent the intellectual creation of the author, which assumes that a person needs to create something. This standard is inconsistent with Section 9(3)49 in terms of doctrine: on the one hand, originality requires the involvement of human intellectual activity, but on the other hand, Section 9(3) permits authorship, which may not involve human intellectual activity.
This conflict can be interpreted as indicating that the UK framework does not emphasize conceptual coherence as much as it should, but rather, it focuses on the functional division of rights. The clause was historically added to deal with the early computer-generated works and to promote the technological advancement, which is rather utilitarian in nature as opposed to the strict observance of the natural law theories of authorship. 43 But when it comes to the contemporary AI, this method runs the risk of over-endowing rights, giving them to the actors whose input is not creative, but infrastructural. This ambiguity is further compounded by the lack of important
40 Copyright, Designs and Patents Act 1988, s 9(3).
41 Toby Bond and Sarah Blair, 'Artificial Intelligence & Copyright: Section 9(3) or Authorship without an Author' (2019) 14 Journal of Intellectual Property Law & Practice 423 DOI: 10.1093/jiplp/jpz056 accessed 17 March 2026.
42 Infopaq International A/S v Danske Dagblades Forening [2009] ECR I-6569
49 Copyright, Designs and Patents Act 1988 (n 47) s 9(3).
43 Toby Bond and Sarah Blair, 'Artificial Intelligence & Copyright: Section 9(3) or Authorship without an Author' (2019) 14 Journal of Intellectual Property Law & Practice 423 DOI: 10.1093/jiplp/jpz056 accessed 17 March 2026.
judicial interpretation such that the practical interpretation of Section 9(3) remains mostly unclear. In this way, though the UK model provides a realistic approach to the issue of ownership, it does this at the expense of conceptual clarity, in terms of which legal fiction is used to fill a conceptual gap that is not yet closed.44
5.2 CHINA
China is a very contrasting case, where judicial innovation and pragmatism are used to address issues related to copyright law, rather than statutory reform. There is a growing recognition of AI-generated works by the Chinese courts as copyrightable, as long as one can prove that there is a human intellectual contribution. The landmark case of Li v. Liu (2023)45 is an example in which the Beijing Internet Court ruled that an image generated by AI could be copyrighted since it was an intellectual effort of the plaintiff to choose prompts, change the parameters, and refine the results. The court pointed out that the work was not a pure machine work but a human choice in terms of creativity and thus met the criterion of originality according to the Chinese copyright law.
46 More importantly, the court disapproved of the idea of AI as an author by reconsidering that authorship should still be considered vested in natural persons, even in technologically mediated situations. Under this model a subtle division of authorship can be made, with the user being recognized as the author in those instances where their effort is substantive and not granted protection in those instances where their effort is minimal or entirely automated. Previous cases demonstrate this changing jurisprudence. The Beijing Internet Court ruled in 2019 that full automated outputs were not eligible to copyright protection, but later cases, such as Li v. Liu,47 show that works which are AI-assisted, but where it is clear that human input has been used, are capable of copyright protection.
48 This development emphasizes the flexibility of the Chinese courts in adapting to technological changes without necessarily having to enact new laws to address the changes. Doctrinally, the Chinese approach is more personhood and labor theory based as it bases authorship on objective human labor and innovative decision-making. Meanwhile, it has utilitarian purposes, 44 Söğüt Atilla, 'Dealing with AI-Generated Works: Lessons from the CDPA Section 9(3)' (2024) 19 Journal of Intellectual Property Law & Practice 43 DOI: 10. 1093/jiplp/jpad102 accessed 3 April 2026.
45 Li Yunkai v Liu Yuanchun (2023) Jing 0491 Min Chu No 11279 (Beijing Internet Court, 2023).
46 Meijian Zhang and Liang Qiang, 'Recognising Copyright of AI-Generated Works' (Law.asia, 8 April 2025 accessed 17 March 2026.
47 ibid 43
48 Yong Wan and Hongxuyang Lu, 'Copyright Protection for AI-Generated Outputs: The Experience from China' (2021) 42 Computer Law & Security Review 105581DOI: 10.1016/j.clsr.2021.105581 accessed 17 March 2026.
promoting the creation of AI technologies and their application by promoting the identification of outputs that have an economic value. Nevertheless, this flexibility is associated with difficulties. Lack of statutory direction can result in inconsistency since courts have wide discretion in deciding on the adequacy of human contribution. However, the model of China is a strong alternative to strict forms of statutory frameworks, as the substantive human active participation is given priority over the formalistic attribution, and therefore, a more logically consistent answer to the problem of AI-generated creations.
5.3 UKRAINE
The Ukrainian law, in a doctrinal way, still has a high human-centric orientation, which means that the subject of the copyright protection must be the result of human intellectual work. This is in line with the standards set by the European Union which focus on originality as one that is the intellectual creation of the author himself. Therefore, it is hard to expect that any pure AI-generated works with no substantial human input will be covered by the existing Ukrainian law as a copyrighted work. Meanwhile, the 2022 law also proposes the idea of non-original objects created by computer programs, which can be granted a sort of a related or sui generis protection, unlike the traditional copyright.
49 It is an effort to make the economic worth of AIgenerated content and the doctrinal need to have human authors without necessitating the total expansion of copyright protection to non-human works. Compared to the silence of India and the official recognition of the UK, Ukraine is placed in the middle ground. It also understands the necessity of legal adaptation but has not yet developed a consistent doctrine of authorship and ownership. Lack of clear attribution remains a problem in terms of enforcement and commercialization, and the use of new legislative constructs creates ambiguity with respect to the extent and type of protection.
However, the changing structure of Ukraine points to a significant trend in international intellectual property legislation the progressive movement towards hybrid and sui generis frameworks that will attempt to support AI-generated works without compromising the central tenet of human authorship. This strategy reflects wider European debates about establishing dedicated rights to AI-generated content, with the goals of innovation and the integrity of the doctrines.
49 Law of Ukraine No 2811-IX on Copyright and Related Rights (1 December 2022)
accessed 3 April 2026.
CONCLUSION
This paper has discussed the central question of whether creations created by AI should be considered property according to the current legal regulations and reached a conclusion that this is possible if interpreted purposively until reforms are implemented. The paper also emphasizes on the practical requirement of recognition. The analysis of classical theories of property, including Lockean Labor theory, utilitarian justification, and personhood theory demonstrate that they are flexible enough to accept AI-generated creation. Although AI systems produce results automatically, the processes behind those are entrenched in human intellectual input, creative guidance, and economic input. The normative bases of property recognition are met by these human contributions.
The silence of the Indian legal system, and especially that of the ‘Transfer of Property Act’, 1882, the ‘Copyright Act’, 1957 and the ‘Patents Act’, 1970, does not constitute exclusion, but indicates a doctrinal gap that can be filled by purposive interpretation. Nevertheless, there is still no clear statutory transparency to bring some clarity in ownership, enforcement, and commercialization, which points to the necessity of legislative intervention in the future. Considering the multi-layered labor involved in AI generated work, the paper argues for join-ownership as the most feasible solution which addresses the contribution of developers and the creative application of users. Arguing for AIcreated works to be considered property is not a change in existing legal standards but a logical extension thereof.
As the boundaries of creativity and production continue to be transformed by technology, property law must evolve to be relevant and in line with the current times and both legal certainty and additional promotion of innovation should be preserved.