- ABSTRACT
The fast development of artificial intelligence (AI) poses new questions in intellectual property law, which is especially problematic concerning inventorship by humans rather than machines and whether AI creations could ever be considered patentable inventions. Traditional patent regimes expect inventorship from a human person, which raises questions about whether autonomous inventions created by machines can lead to patent protection. This article examines the legal, ethical, and practical implications of AI-generated works, drawing on several recent cases as well as our domestic and international statutory language. We examine the criteria for patentability and, more broadly, patent eligibility; including the novelty, inventive step, and industrial applicability, and whether these can work in claims when inventions are prepared, created, or invented by machines. In addition, the above discussion leads to issues around ownership, specifically who should hold patent rights should they be the AI developer, the User, or the machine itself? Through a critical evaluation of emerging trends and jurisdictional responses to patenting machine inventions, this article provides a clearer picture of the ways we might modernize our intellectual property systems to accommodate technological impetus, all whilst balancing incentivising creators, legal certainty and making sure that the public interests are considered.
KEYWORDS
AI-generated inventions, patentability, machine-created works, intellectual property, ownership, legal frameworks
- INTRODUCTION
The rise of Artificial Intelligence (AI) has transformed the innovation landscape by providing machines the ability to independently create inventions, designs, and other works. This capability disrupts existing intellectual property (IP) paradigms that have historically been anchored in human authorship and inventorship. The patent system, in particular, relies on the idea that inventors are natural persons capable of exercising judgment and creativity in ways that AI systems generally do not or cannot meet. An emerging issue is the extent to which patents can cover inventions created by AI and, consequently, the ownership of related rights whenever AI is involved. Different jurisdictions around the world have begun to address innovative machine inventions and the extent of patent protection that will apply to rights produced by AI. For example, the United States, United Kingdom, and European Patent Office have taken different approaches as to whether AI can be recognized as an inventor. In the debate surrounding AI-generated inventions, ethical and economic considerations will emerge, including: how AI drives market-based incentives for innovation; how to fairly attribute rights; and the social consequences of enabling machines to create without human involvement. This paper will analyse the intersection of IP law and AI and analyse statutes, case law, and policy debates to evaluate the extent to which existing legal systems can accommodate AI-generated inventions, while respecting the underlying reasons for the existence of patent protection.
- RESEARCH METHODOLOGY
This research takes a doctrinal and analytical approach to investigate the patentability of inventions generated by AI and the ownership of works generated by machines. The main primary sources utilized in this research study are national and international IP legislation, patent legislation, and case law from various jurisdictions, including the US, UK, European Patent Office, and India. There is secondary source material in the form of academic articles, reports, and policy papers pertaining to AI and IP. This research study also included a comparison of how various jurisdictions handle the issue of AI inventorship and ownership. The implications of this research are discussed from legal, ethical, and economic perspectives, identifying existing holes in previous legislation. The research involves qualitative data collection of legal texts, judicial interpretations, and comments and opinions by experts in the field. Given these methodologies, this study provides a complete overview of AI, while studying its interface with IP law, which will provide recommendations for modifying existing legal regimes in response to technological innovation.
- REVIEW OF LITERATURE
The increasing discourse surrounding AI-generated inventions reveals considerable challenges within patent law and ownership. Abbott (2020) (Abbott, 2020) points out that current patent laws are unprepared to accommodate non-human inventors and is calling for changes in legal structures to ensure that AI-fuelled inventions are able to contribute to further innovation. Thaler’s (2021) (S., 2021) investigations provide instances in which AI’s creation elements resulted in patentable inventions, producing different arguments regarding authorship and ownership. The European Patent Office reports (2022) (Office, 2022) illustrate an uneasiness developing among jurisdictions related to whether AI should be recognized as an inventor, when situations of non-human inventorship arise, but these discussions do not appear to have implanted roots in the US as the discussions continue to advance and advocate for a broader interpretation of the scope of inventorship. Kapoor and Singh (2023) (Kapoor, 2023) provide an examination of ethical and economic issues with AI-generated invention and emphasize the importance of clarity among stakeholders so that a balance can be struck between stimulating innovation and public interests. Finally, WIPO’s (2021) (Organization, 2021) comparative international survey of AI-generated invention related works of authorship reveals inconsistency regarding attribution, illustrating the urgency of cohesion among intellectual property laws on a global scale.
- PATENTABILITY OF AI-GENERATED INVENTIONS
Being patentable or not, AI-generated inventions pose perhaps the most complex set of technology and law issues that go against the grain of those frameworks that insisted on human inventorship. The section focuses on the dynamic intricacies of the law and highlights important case laws and statutory provisions and judicial interpretations across jurisdictions.
- United States
In the landmark case Thaler v. Vidal (2022) (Thaler v. Vidal, 2022), the U.S. Federal Circuit upheld the Patent and Trademark Office’s (USPTO’s) position that pursuant to the Patent Act, inventors must be natural persons. The court found “inventor” to mean a human being, thus precluding an AI system from being an inventor. The United States Supreme Court then declined to review the case, effectively resolving the matter in favour of the Federal Circuit.
- United Kingdom
In a similar vein, the UK Supreme Court set aside recognition of inventive capacity upon AI. Dr. Stephen Thaler tried to have his AI system, DABUS, named as the inventor on patent applications. The Court emphasized that the statutory language of the Patents Act 1977 requires an inventor to be a person, thereby excluding AI systems from inventorship.
- European Union
The EPO, has repeatedly asserted that only natural persons can serve as inventors. In 2022 the EPO Board of Appeal rejected an application listing DABUS as the inventor further supporting that inventorship is a human idea.
- Germany
On the other hand, in 2024, Germany’s Federal Court of Justice (FCJ) broke ground, ruling that patents for inventions produced by AI were permissible. The court ruled that although AI can create inventions, if a human holds a patent, the human must be an inventor if they substantially prepared, or directed the output of, the AI system to come to an applicable invention. This approach is in keeping with current tests for patent protection under German law.
- India
While India’s patent law is found in the Patents Act, 1970, it is silent as to the patentability of AI-generated inventions. In fact, in recent applications, the Indian Patent Office has objected to the recognition of AI as an inventor, citing that an inventor must be a “true and first inventor” which implies human inventorship.
- International Perspectives
The international debate rages on, with countries taking divergent approaches to AI and inventorship. For example, South Africa granted several patents, which listed an AI as the inventor, but only on procedural grounds. In other jurisdictions, like Switzerland, they have upheld human inventorship, following their existing practices.
- OWNERSHIP OF MACHINE-CREATED WORKS: NAVIGATING THE LEGAL LANDSCAPE
The ownership of works produced by artificial intelligence (AI) in India is a complex and evolving issue primarily because there are currently no express provision for the status of AI as an author under the 1957 Copyright Act. Section 2(d) (Indian Copyright Act, 1957) of the Act describes an author as a person who makes a work; it does not make provisions for non-human (editors) authors. These gaps leave ambiguous the question of the position of works of AI. Where things become murky is whether AI can even be regarded as an author, or whether ownership should be presumed to rest with the persons who created the AI or the person who operates it.
India has recently had a few occurrences of AI systems being recognized as joint authorship and entitled to rights along with human authors. For example, a painting called “Suryast” was created by an AI application called Raghav. Several applications emerged claiming that the work should have joint authorship between the AI application and the programmer of Raghav. Interestingly, although this is a forward-thinking notion, it begs the question as to how much human contribution must exist in order for joint authorship to qualify under current law. The position of the government, as per published comments made by Union Minister Shri Som Parkash, seems to be that current laws are sufficient to answer AI-derived works, although as AI capability develops the legitimacy of their position is being increasingly challenged.
The approaches toward AI-generated works vary across the globe. In the United States, the Copyright Office has denied registration for works created only by AI on the grounds that human authorship is required under the Copyright Statute. Similarly, the EU’s obligations and the UK statutes have established that human creation is necessary for copyright. This makes it abundantly clear that while countries share a disposition, they have unique legal traditions that establish challenges for India in a global perspective.
Currently, there is ambiguity in the law regarding the ownership of AI-generated works. This ambiguity could lead to disputes. As AI continues to rapidly develop and becomes more sophisticated in generating creative processes, there is way for legislative amendments to definitively designate what role AI has as an author and establish ownership rights for AI-generated works. Legislative revisions would also clarity in the law and promote innovation by giving creators, contributors, and developers acknowledged ownership rights.
In short, India has begun accepting AI’s possibility in the creative landscape, but the outline of AI-generated work’s ownership still lacks sophistication in the legal realm. To create a fertile scenario for innovation and protect all stakeholders involved, Indian copyright law must evolve and explicitly incorporate the complexities of authorship and ownership now offered by AI.
- ETHICAL AND ECONOMIC IMPLICATIONS OF AI-GENERATED WORKS
The rise of artificial intelligence (AI) in creative fields has given rise to new ethical and economic dilemmas, especially on issues of ownership and monetisation of AI-generated works.
- Ethical Considerations
One of the key ethical issues is the question of authorship. Traditional intellectual property (IP) laws, such as the Indian Copyright Act, 1957, require a human author for a work to receive protection. Section 2(d) states that an “author” is a person that makes a work, consequently, AI systems cannot be considered as authors. This lack of authorship raises issues about fairness, as works have substantial human input in the programming and training of AI systems, so denying these works protection which are AI generated may not be fair.
Also, recognizing AI as inventor or author raises ethical issues about accountability, focusing on the moral person and dilution of human creativity. If rights are assigned to AI, this could devalue human contributions and change the traditional concept of creativity and innovation itself.
- Economic Implications
From an economic perspective, AI-generated works offer the potential for gains and loss. As an example, AI can reduce creation costs and increase the ability for individuals to access new creative technologies. However, it is also possible that no matter the price of the content, AI works will damage the market for human-created works. For instance, the International Confederation of Societies of Authors and Composers (CISAC) estimates that musicians will lose 24% of income from AI by 2028 or €22 billion, as human compositions will be in competition with AI compositions.
Additionally, using copyrighted material to train AI without licensing and compensating the original creators creates a question of fair use and exploitation by the creators of AI. Law suits against technology companies like Meta, Stability AI and GitHub represent the growing tension between progress in technology and protection of creator rights.
- Legal Framework and Policy Considerations
The gaps in specific legislation governing AI-generated works presents a pressing need for policy responses. In India, the Copyright Act does not accommodate for non-human authors and makes it unclear whether AI-generated content is protected at all. At the international level, jurisdictions vary in where they land on this issue, from adapting their IP laws recognizing the role of AI in the creative process, to jurisdictions that still have a hard requirement for the authorship of an industry practitioner.
India may consider amendments to the Copyright Act by introducing legislation that creates policy for AI-generated works, so that creators who were employed building the AI are recognized and compensated for their work. The importance of reforming policy in this area cannot be overstated, as the state of Copyright would otherwise remain stagnant, in a Global situation where the recognition of authorship by AI is gaining momentum and supporting creative industry development in the AI age.
The ethical and economic considerations related to AI-generated works underscore the need for a neutral approach which enables individuals to innovate and explore new creations as well as protect human creators’ interests. Law is necessary to mitigate the unique features that AI introduces into the creative sector.
- POLICY AND REFORM RECOMMENDATIONS FOR AI-GENERATED WORKS IN INDIA
The existing laws regarding copyright in India, primarily the Copyright Act of 1957, currently do not consider the complexities that artificial intelligence (AI) has introduced into the production of creative works. At the present, this absence of statutory consideration, leads to ambiguities and has raised questions regarding copyright and protection of works created with the assistance of A.I. In order to solve and move past these issues, and to encourage innovation, we suggest the following policy changes:
- Defining AI-Generated Works
In consideration of a technological shift, the Copyright Act should be adjusted to provide an explicit definition of AI-generated works. If we look at the UK’s approach, which attributes authorship to the person who makes arrangements for the creation of a work, perhaps India could accommodate authorship to the individual or entity that supplies the inputs or instructions to the AI system, in order make clear that AI-generated content is included in the eyes of the law, consequently providing creators clarity and protection.
- Introducing Fair Use Guidelines for AI Training
The existing legal structure has no explicit rules regarding the applicability of copyrighted material for training AI systems. India should implement conditional fair dealing law provisions supporting AI training on copyrighted material in datasets that are legally acquired and properly attributed, to balance the rights of originators and innovation. This would reflect an internationally understood practice and facilitate ethical AI.
- Establishing Liability and Accountability Mechanisms
The move towards greater autonomy of AI systems complicates the attribution of liability for infringements, which call for clearer laws to assign accountability to developers, users and other stakeholders. It is reasonable to require this as we grapple with institutional culpability of AI systems, and to require both provisions for attribution of liability for any AI produced content not infringing existing copyrights, as well as redress for any harms.
- International Harmonization
Given that AI is part of a global ecosystem of technology, India should be involved in international discussions to harmonize copyright laws regarding AI-generated material. Involving other countries will help them develop a cohesive set of laws that deal with the challenges AI raises for the creative industries and prevent its creators and developers from being affected in the global creative marketplace.
- Public Awareness and Capacity Building
For these reforms to be successful there will need to be a significant increase in awareness and knowledge about the impacts of this technology on copyright amongst stakeholders. This could involve educating creators, developers, legal practitioners and policymakers (in particular) on the rapidly changing landscape and opportunity to rethink legal frameworks to update them to reflect changing technology. Through capacity building initiatives, all stakeholders would be better equipped with the knowledge and skills to understand how to manage AI-generated works.
- SUGGESTIONS
Imposing legal, ethical or economic approaches globally to deal with the consequences of AI-produced works, could be treated in a multitude of ways. First, amend Indian intellectual property law to specifically recognize AI-assisted works, or machine-generated works. Relevant sections, for example, section 2(d) of the Copyright Act, 1957 and relevant sections of the Patents Act, 1970, can be amended to reflect the issue of authorship or inventorship and refer to the human operators involved, who are training and/or supervising AI systems.
Second, establish conditional fair use or fair dealing for datasets used within AI-training datasets. This would allow copyright material, derived as output of an AI model, protected works, while accomplishing the objective of fairy compensating original users/producers of the material. This would assist in the innovation of new works while promoting fairness in compensating copyright owners.
Third, a framework of attribution for ownership can be established. The rights can vest with AI developers, programmers or users, as identified by their contributions to the output generated by the AI. This process would limit ownership arguments and drive clarity for commercialization of outputs produced by AI.
Fourthly, legal obligations of liability and accountability should be consolidated. Clear definitions of liability should ascertain who would be responsible for copyright infringement or any other misuse of AI-generated material or output. AI must be developed and promoted ethically, with avenues to seek legal remedies available to harmed parties.
Fifth, international harmonization and policy discussion should be emphasized. To ensure international consistency, it is crucial to align India’s IP laws, to the best extent possible, with global standards. This will avert potential jurisdictional discrepancies and cross-border commercialization of AI innovations.
Finally, raising awareness and capacity building is vital. All stakeholders, including creators, developers, and policy makers need to be informed about the consequences AI has on intellectual property, the ethical implications, and best practices associated with its adoption.
- CONCLUSION
Artificial intelligence has developed into an agent of change in innovation and creativity, but its evolution is exposing significant gaps in existing intellectual property frameworks. Current laws in India and globally, largely endorse human inventorship and authorship, leaving questions of patentability and ownership of AI-driven creations uncertain. Existing case law like Thaler v. Vidal and case law from the UK, EU and India contend with the tension between existing legal frameworks and disruptive technology. Moral issues from ethical attribution and accountability to the economic impact on human creators further amplify the need for reform. The policy measures recommended in our paper such as amendments to the Copyright Act and Patents Act to enable AI driven creation; including fair use provisions for AI Training; clarifying ownership frameworks; and accommodating international harmonization are intended to help reconcile innovation with fairness. To protect human creativity, ensure incentives for innovation and allow India to continue to compete in the global technological landscape, legal regimes must be modified to address AI driven creation.
REFERENCE
- Abbott, R. (2020). AI and the Patent System: Adapting Law for Non-Human Inventors.
- Indian Copyright Act, 1. (1957). § 2(d), No. 14, Acts of Parliament. India.
- Kapoor, A. &. (2023). Ethical and Economic Considerations of AI-Generated Inventions.
- Office, E. P. (2022). Patentability and Inventorship in AI-Generated Inventions: Annual Report.
- Organization, W. I. (2021). Comparative Study on AI-Generated Inventions and Authorship.
- S., T. (2021). Artificial Intelligence as Inventor: Challenges for Patent Law.
- Thaler v. Vidal, 21-2347 (United States Court of Appeals for the Federal Circuit August 5, 2022).
Submitted by Praneetha Godula, student Symbiosis Law School, Hyderabad.