top of page
Search

AI-GENERATED SCIENTIFIC RESEARCH AND PATENTABILITY: RETHINKING INVENTORSHIP IN THE AGE OF AUTOMATION

  • sonalimukhia2002
  • Jul 31
  • 15 min read

Author: Mitushi Kankerwal, Symbiosis Law School, Hyderabad


ABSTRACT

Artificial Intelligence (AI) has revolutionized scientific research from manual data analysis to autonomous closed-loop laboratories that can design, conduct and tweak experiments independently. This paradigm shift presents a serious challenge to global IP systems that have been traditionally limiting patent inventorship in ways that are confined to natural persons (human beings). As a result, the right to patent AI-generated research is tricky due to the fact that AI is not considered a legal person that can own a right or carry out a right. The DABUS patent applications clearly showed this split of laws, as South Africa and Australia had decided to accept AI as the inventor, whereas other major patent offices, including the USPTO, UKIPO, the EPO and the Indian Patent Office, have refused to accept AI as the inventor on the basis of the doctrine of human exclusivity. As an example, under the Indian Patent Act of 1970, it is a requirement that the inventor must be a "person," which is defined as a natural or legal person, not AI systems. If protection of AI-generated inventions is not adequately achieved, the paper cautions, it could lead to an "innovation vacuum," in which corporations may have to conceal technological advances as “temporary trade secrets” instead of publishing them to the world, thus violating the social contract on which patent law is built. The researchers suggest that this "deadlock" be addressed by a move to an "augmented inventorship" framework, to be accomplished through deliberate legislative action instead of case-by-case judicial decisions. In this model, the human users who offer much original direction and feedback to the AI and interact with its useful “system layer” would have patent rights, with AI clearly designated as a sophisticated tool. The modernization of statutes such as Section 6 of the Indian Patents Act, through a participatory process involving all stakeholders, is crucial to safeguard the economic interest of human beings, to ensure the sharing of knowledge will not cease, and to optimize the many benefits of automated innovation without risk.


KEYWORDS

Artificial Intelligence, Innovation, Patent law, Technology, Intellectual Property Rights, Inventorship.


INTRODUCTION

Unlike traditional research which includes manual analysis and executes human instructions at better speed, autonomous laboratories utilize machine learning and represent a major shift from simple automation to closed loop in scientific research; they independently design, perform, and adjust experiments. The rise of generative ai platforms which helps in designing novel chemical compounds in minutes. Artificial Intelligence plays a crucial role in innovation and technology. Earlier the things which used to take a lot of time are now with the help of artificial intelligence it can be done in minutes. Artificial Intelligence is taking over a lot of jobs whether it is law, medicine or any industry. While intellectual property (IP) laws globally still restrict patents to human beings who invent them but not AI. If an AI innovates, you cannot give the patent to AI itself, law still recognizes the legal human owner of the AI to usually claim the patents. Earlier lawyers used to take a lot of time to check if an invention is already existing but now with AI tools, it can search hundreds of documents globally in very less time to check if an idea is new. Patents are rights of intellectual property. Patents are territorial, meaning they offer protection only in the jurisdiction where an application has been filed and a patent granted.


LITERATURE REVIEW

A patent is a type of IP which gives the holder exclusive rights to the innovation. The issuance of a patent is subject to the following requirements like patentable subject matter which identifies the type of works that can be accepted as an invention. The courts have held that laws of nature, physical occurrences and abstract concepts are not patentable. The Supreme Court of the United States has said, “Congress intended to bring into the tax base everything under the sun that is made by man.” The second condition is to find out whether such an innovation has a beneficial purpose and the invention must have a practical usage and the invention must not be known or used by others and it must be a new product or process. The subject matter should not be obvious to a person having ordinary skill in the art.

Section 2 and Section 6 of the Indian Patent Act, 1970 elaborates on the criteria for recognising an inventor and the person who can actually file for a patent in the Indian jurisdiction. In particular, Section 6 gives details about who is considered a person entitled to apply for a patent. Section 2(1)(s) basically describes who can be referred to as a ‘Person’ on the other hand, Section 2(1)(y) discloses the persons who, when it comes to filing a patent application, cannot be treated as a true and first inventor.

Section 2(1)(s) of the Act clearly says that the person filing for a patent can be either a “natural person” or any “Government” organization. Still, it is interesting to see how the term “person” is not just limited to a natural person and/or Government, because it also seems to bring in Government together with other entities that may be able to claim they are the real and first inventor of the invention. In that sense, it can be construed that anybody, or anything, that is the first and true inventor of the invention qualifies as the “person” for whom the patent application is being submitted . The USPTO Inventorship Guidance for AI-assisted inventions is being revised so that only natural human beings can be named inventors on a patent. AI systems can’t be listed as inventors or as co-inventors but that doesn’t mean people lose everything. Human inventors can still lean on AI as a tool, without it automatically taking away their patent rights,as long as they made a significant original contribution to the way the invention was created. In the case of the Thaler v. Comptroller General of Patents, Designs and Trademarks The issue was brought before the Court after some granting of patents for inventions that were made by machine. This matter actually concerns two British patent applications for two separate inventions, Dr. Thaler mentioned in his application that were generated by an AI machine called DABUS. Under Section 13(2) of the 1977 Act, the UK Intellectual Property Office, usually shortened as UKIPO, refused the applications because DABUS isn’t a person, as that term is meant in sections 7 and 13 of the 1977 Act. After that, UKIPO’s decision was taken up by appeal before the High Court, and also before the Court of Appeal at the same time, but in the end both appeals were dismissed. Then, when it finally reached the Supreme Court, the Court, while handling this kind of question, has explicitly stated that an apparatus, or a machine devised product or procedure, cannot be patented.

Inventions based on AI have critical layers of data, application and system. Under the eligibility framework of AI, the data and application layers are not deemed to be such a technical part of the invention as they pertain to merely abstract notions. So, what may be patented? The answer is the system layer which is called the user-interaction layer of AI-based technology which runs programs, displays output etc. In simple terms, a system which has physical or practical importance is patentable. Therefore, it is important to substantially incorporate key components or features of AI in a practical aspect so as for AI-like technologies to qualify as patentable subject matter and avoid falling under excluded subject matter. Patent laws globally require a natural person to be an inventor, if an AI is doing the heavy lifting, who will get the patent?


METHODOLOGY

The research methodology is doctrinal in nature with an analytical research which includes analysis of statues, cases and regulations to provide a structured approach for rethinking inventorship in AI generated scientific research and patentability. This approach will help in clearly analysing and helping with understanding of patentability in AI generated research. The patentability framework of various countries such as the USA, UK and Japan is also discussed in this research paper. There is no survey or case study work, and the information has been obtained from books, journals, articles, research papers, publications and internet databases. The researcher has extensively consulted legal databases that are reliable to gain an understanding of the legal landscape around Generative AI and to develop objectives that can serve as the foundation for research. Globally, the rights of patents are protected for humans, but in this paper there will be the need for an ‘augmented inventorship' due to autonomous discovery so as to avoid an ‘innovation vacuum’.


THE IMPLICATIONS OF ARTIFICIAL INTELLIGENCE AND INTELLECTUAL PROPERTY LAWS

The advancement of artificial intelligence applications and inventions is strongly linked to inventorship. The idea behind the patent system is to provide the inventor with certain rights in the invention. If it is not known who created the patent, then the patent may be invalid. The issue with AI is that there's an unclear debate about whether an AI can be the inventor and who the actual inventor is, as for the patent rights. In different jurisdictions, the meaning of inventor could vary but is usually understood to refer to an individual. In the U.S., the court case “Diamond v. Chakrabarty” widened the definition of what can be considered an invention and made whatever produced by a human patentable. The idea is to retain the concept within the individual (not in a formal organisation such as a company). Inventions made by an AI is a complex issue because it is not recognized as a legal person and thus giving ownership of patent rights becomes complicated as AI cannot own or execute rights independently.Absolute liability holds the owner responsible for any infringement done by AI and identifying the violator is important for both the patent holder’s right and the common world. The present patent system does not consider AI as infringers, thus the user or the owner of the AI generated research would probably be held accountable. In today's world patentability on AI’s work requires careful consideration. Artificial intelligence in earlier times used to simplify the work load and helps in reducing the involvement of natural persons but nowadays the roles are reversed, AI is taking away jobs and innovating new ideas and codes. The definition of an “inventor” in patent law does not include AI-generated research and needs to be thought over in the light of AI in innovation.


GLOBAL LEGAL STANCE

The European Union is encouraging other countries to include AI related work in innovations as “own intellectual production.” Patent protection should also be considered for AI-generated research. It's a good sign for the recognition of AI's distinctiveness in creating artistic works. AI and robot innovations are also subject to a high degree of autonomy and can execute tasks without human interference; therefore, they should also be protected by patents. Patent protection is more complicated for concepts generated by AI, however, because they must satisfy the three part test of novelty, distinctiveness and industrial applicability. Additional research and clarity of existing legislation/regulations is required to address patentability and other issues of AI based concepts. In addition, the report of the Ayyangar Committee 1959 reveals the legislative purpose for enacting the Patent Act, which intends to mention an inventor as a right of inventor for a relevant patent. The report emphasizes that any person who has a moral right to be named as an inventor, even though he might not have all the legal rights over the invention. According to the report, the idea of approaching an inventor about their creation is to help the inventors increase the financial value to which they are legally entitled, even if they may have given up their exclusive rights through contracts or agreements. But the AI is not going to have the luxury of the benefits promised by the legislative intent nor will it give the moral rights accorded by existing Indian laws nor will it be able to stop a person from misusing it. The current international, Indian and various national regulations make it abundantly clear that the inventor is required to be a natural person. In this respect it may be considered to what extent such a provision will fulfil the main function of the patent system in the social and economic situation of the immediate future. The purpose of the patent system is to spread the knowledge and usefulness of the invention to the whole of society through compulsory publication and to ensure that the knowledge is spread and that it is useful to society; technology is to be developed further. Under certain conditions, the right holder is given an exclusive right in the market over that patented invention for a limited duration, thus giving him an economic interest attached to it. Hence, it is important to consider what if in the future inventions made with the help of AI are not covered by the protection. However, there is a high risk that a lot of knowledge can remain a trade secret for AI-generated inventions, which could have the effect of slowing down the development of technology.

The European Union received two patent applications with DABUS listed as an inventor in the year 2019, DABUS (Device for the Autonomous Bootstrapping of Unified Sentience) was claimed in the name of a person named Dr. Stephen Thaler. It was listed as an inventor on the patent list because it claimed to be true with two discrete inventions without any human intervention. A lot of legal challenges and global intellectual property concerns are raised to assign inventorship to an AI-machine and whether it can be named as the inventor of AI-related patents. There are only two jurisdictions where the DABUS case was filed and have recognised and accepted it as an inventor i.e South African Patent Office and Federal Court of Australia, whereas other countries did not recognise AI as an inventor and asserted it should be a natural person.

In South Africa the South African Patent Office was the first IP office to grant a patent for an invention created by the AI machine DABUS. It is important to note however, that South African patent law does not include the term “inventor”. On July 2021, the Australian Federal Court ruled that AI machines could be inventors, overturning the Australian Patent Office's decision to refuse to allow them to be an inventor. The reasons for this decision included that an agent might be an inventing person or thing, with the agent defined as computer, regulator, dishwasher and like there; and there was no provision in the Australian Patent Act that barred an AI system from being an inventor. The court in its judgement cleared that the ruling affects only to the inventorship of patents and not the ownership.

The European Patent Office (EPO) and United Kingdom Intellectual Patent Office (UKIPO) have both denied inventorship status for an artificial intelligence (AI) machine, unlike South Africa and Australia. The United States Patent and Trademark Office (USPTO) also rejected DABUS reasoning that a plain reading of statutory provisions of United States Code (U.S.C.) discloses the inventors only as natural persons. In reference to the inventorship of an invention, the USPTO referred to the provisions of 35 U.S.C. Sections 100(f)-(g) and Section 101, which use the term “individual,” and “whoever” in that context. A declaration or oath of invention is to be signed by a person who considers himself or herself to be the inventor of the invention claimed in the patent application. In addition, the USPTO referenced a number of Federal Circuit cases that held that only natural persons could be inventors. The EPO concluded that the term "inventor" in the European Patent Convention only applies to a natural person. Not every title, such as that of "inventor", is a legal status which can be exercised by a machine, the EPO also emphasized, and merely assigning a name to a machine does not make it a legal entity. The UKIPO has made its decision on the grounds of arguments put forward by the EPO.

Thus, courts in the USA, UK, and EU have rejected AI patent applications, emphasizing the requirement for human inventors. Japan has adopted AI policies to improve outcomes and ensure that AI-related patent applications are acceptable, which necessitates proving that AI-generated inventions do not form part of prior art. In contrast, in India, the DABUS patent application was objected to by the Controller General of Patents because AI inventions are not acknowledged under Section 2 and Section 6 of the Patent Act, 1970, as DABUS is not recognized as a person in India.The legal structure in India is based on the Indian Patents Act, 1970, which mandates that a patent application must be submitted by a "person" who is the "true and first" inventor or their assignee, as outlined in Section 6 of the Act. The General Clauses Act of 1897 defines a "person" as a natural or legal entity, excluding AI systems.Section 2(y) of the Act defines a "patentee" as a person who obtains a patent. According to Section 2(1)(j), an invention is defined as a "new product or process which has an inventive step and is capable of industrial application." The inventive step necessitates human creativity, which further reinforces the argument that AI cannot be patented under the current legal system. This viewpoint was supported in the case of V.B.Mohammed Ibrahim v. Alfred Schafranek.


ANALYSIS

The future of the patent system with AI is bright and full of potential, but also fraught with challenges. As AI, the more it evolves and the more it creates complex and innovative inventions. The current patent system will have to evolve adequately to compensate and reward these advancements. The questions presented by the USPTO in their call for comments span across many questions which includes inventorship related issues, such as involvement of AI in the process of creating an invention, and ownership and patentability of AI-generated inventions. It is an important step towards setting a model for the development of IP law. USPTO is seeking input from experts and stakeholders to inform its decision-making process. It seeks advice from experts and stakeholders for its decision-making process and India too can benefit from it by following a similar path and engaging with its own experts and stakeholders to respond to issues of who can own an invention created by AI. This would enable the Indian government to develop a more extensive approach to handling challenges of AI-generated Inventions, assist in the development of and ensure the ongoing role of intellectual property law in encouraging innovation and public benefit. As of now, under the Indian Patent Act, only a "person" can be named and apply for a patent as an 'inventor'. It is not clear, therefore, whether AI can be said to be truly artificial intelligence. As per Indian law, an inventor is someone who invents an innovation and it would be better to revise Patent Law to expressly include the problem of AI inventorship. This may include the establishment of a new category of inventorship to address the distinct nature of AI, through AI systems, or by adjusting current legal frameworks to consider the specific nature of AI technology. As AI technology evolves rapidly, it is essential to have a consistent strategy to manage AI. However, this should be done not by judicial means but by the adoption of legal reform intervention.


CONCLUSION

The shift of Artificial Intelligence (AI) from an automated tool to a fully-fledged creator is a fundamental paradigm change in the process of scientific discovery. The current IP system is undergoing an unprecedented crisis from its foundations as shown in this paper. What is fundamentally missing from the current legal framework, is the ability to manage generative platforms that can independently design, execute and optimize complex innovations within minutes.

A survey of the world's exploration of AI inventorship shows a highly fractured legal landscape. One side has firm advocates for doctrine of human exclusivity such as USPTO, UKIPO, EPO and India and on the other side has very strict advocates for the doctrine of human exclusivity in the United States (USPTO), United Kingdom (UKIPO), European Union (EPO) and India. These regimes are based on the statutory requirements and on pivotal decisions such as the DABUS rejections, and the India case of V.B. Mohammed Ibrahim v. Alfred Schafranek, which both consider that an "inventor" must be a natural person. But anomalies in South Africa and early judicial changes in Australia indicate a rapid understanding that the actual use of contemporary innovation exceeds and is outstripping its legal jargon.

But disregarding AI in creation is a huge risk. The scientific world could become an "innovation vacuum" if patent offices refuse to provide protection to what AI has done for the invention. Corporations will of course cease filing for public patents in favor of permanent trade secrets to protect their commercial interests. This change would directly break the social compact that has been fundamental to patent law: public disclosure for a temporary monopoly in the market. If technological knowledge cannot be shared openly, the transfer of knowledge within and across communities will come to a standstill, which will ultimately hurt society as a whole.

This paper proposes that the current system of international IP needs to shift towards an “augmented inventorship” system to overcome this deadlock. This shift should be one in which law-makers rather than judges make the decision for it. Key steps that are essential to the modernisation of the current patent landscape are like the patent evaluation should not just consider the abstract data layers but also the human contribution in the "system layer," which is the interface where humans guide, prompt, and use AI tools in the physical and practical context. AI should be explicitly programmed as a sophisticated tool of human purpose. As long as a natural person gives significant original direction, then patent rights ought to firmly stay with the human proprietor or user. India and other jurisdictions should move forward actively engaging legal experts, scientists and industry stakeholders to systematically modify provisions, like Section 6 of Indian Patents Act, 1970, in a collaborative effort, initiated by the USPTO. To sum up, AI is not just a passive machine that types out what people say; it's a driving force in industry's evolution. An augmented inventorship mechanism will help patent laws to effectively protect human economic interests, respect the historical understanding of who can be a "person" of a legal act, and unlock the potential of AI without any constraints.


REFERENCES

BOOKS

  • Ben Hattenbach & Joshua Glucoft, Patents in an Era of Infinite Monkeys and Artificial Intelligence.

  • James Daily & F. Kieff, Anything Under the Sun Made by Humans: Patent Law Doctrines As Endogenous Institutions for Commercializing Innovation, 62 Emory Law Journal 967 (2013).


JOURNAL ARTICLES

  • Aranya Nath & Gautami Chakravarty, Patent Registration Criteria for AI Inventions: A Critical Analysis of Evolving Standards, IP Bulletin, Vol. V, Issue 1, Jan.–Jun. 2024, at 6 (2024).

  • Dr. Deepak Kumar Srivastava & Aryan Kare, Artificial Intelligence and Patent Law: Analyzing Patentability of AI-Generated Works, 63(1) Panjab University Law Review (2024).

  • Namrata Sahu & Priyesh Ghosh, True and First Inventor: AI-Generated Inventions, 6(3) International Journal of Law Management & Humanities 2782 (2023).


CASE LAWS

  • Diamond v. Chakrabarty, 447 U.S. 303 (1980).

  • Thaler v. Comptroller General of Patents, Designs and Trademarks, [2023] UKSC 49.

  • V.B. Mohammed Ibrahim v. Alfred Schafranek, AIR 1960 Mys. 173.


STATUTES

India

  • Patent Act, 1970, No. 39 of 1970 (India).

  • Patent Act, 1970, §§ 2 & 6 (India).


United States

  • 35 U.S.C. §§ 100–101.


GOVERNMENT / OFFICIAL PUBLICATIONS

  • Federal Register, Request for Comments Regarding Artificial Intelligence and Inventorship (Feb. 14, 2023).

  • United States Patent and Trademark Office, Revised Inventorship Guidance for AI-Assisted Inventions (Nov. 26, 2025).


WEB SOURCES






Comments

bottom of page