AI-Generated Inventions & Patents: Can an AI System be Named as an Inventor Under the Patents Act 1970, or Must Human Inventorship Apply?

Introduction

Artificial intelligence is no longer just aiding in research; it is also providing solutions in the area of drug development, materials development, and engineering. This raises the question of who the inventor in patent law when the machine is responsible for the invention. The question was raised on a global level due to the DABUS case wherein Dr. Stephen Thaler recognized an AI program as the sole inventor of an invention that is a food container and emergency light. The Patents Act, 1970, clear on one aspect but vague on another. AI cannot be considered as the inventor but how it will be classified is still unclear.

Explore More: Patent Law

Inventorship under Patents Act, 1970

The expression “inventor” is not explained by the Act itself. Rather, it is understood through “true and first inventor” which is mentioned in Section 6(1)(a) which permits the application of that inventor, his assignee, or legal representative.

The Act also specifies the meaning of “person” in Section 2(1)(s) and thus includes natural persons, companies, societies, and governmental bodies, however, it excludes software in this definition.

Section 2(1)(j) explains the concept of “invention” as something new that can be either a product or a process and involves the inventive step that is useful for industrial purposes.

Section 2(1)(ja) states that inventive step is something to be considered from the perspective of a person with relevant professional experience in a certain field.

Case Study & Global Perspective: The DABUS Case

In this connection, Dr. Stephen Thaler, a Computer Scientist submitted a patent application on a light source and food container produced by DABUS, an AI System. This whole situation centres around the question of inventorship by machines being possible.

In this regard, the majority of authorities, patent offices among them, rejected the idea about the existence of the inventors that are AI. In the UK, the Supreme Court in Thaler v. Comptroller-General (2023) has ruled that only a human can be an inventor, and that an AI has no possibility of being defined as such. The Federal Circuit of the USA in Thaler v. Vidal (2022) reached a fairly similar conclusion that says that inventor is a defined person from the standpoint of the Patent Act and it does not include AI systems. The Federal Court in Germany while handling a similar case stated that a human being remains the inventor and AI could be given credit for the invention, but not inventorship. Only in South Africa, the DABUS patent was granted in 2021, but without any proper review or examination, which made the decision questionable in the eyes of the legal fraternity.

Section 2(1)(y) of the Patents Act, 1970 defines a “natural person.” And Section 6 states that only such a natural person could be considered an inventor in the eyes of the law. A Declaration of the inventor, as a consequence, must be attached to any application for patent being made. On the basis of this fact, it can be claimed that Indian legislation would require a human inventor and the issue of AI inventions may be addressed by lawmakers.

AI-Generated Inventions in India: Statutory Framework & Legal Position

In the first place, mathematical and business methods, computer programs, and algorithms were expressly excluded from being patentable as per Section 3(k) of the Patents Act, 1970. Any applications of AI which claim a patent and include abstract algorithms are unlikely to succeed.

In Ferid Allani vs. Union of India (2019), the Delhi High Court indicated that patents related to computers must be examined on a substantive basis without caring whether the product involves any program.

In the case of Caleb Suresh Motupalli v. Controller of Patents (2025), the Madras High Court observed that it was the responsibility of the judge to address objections revolving around the lack of clarity and sufficiency of disclosure. It will not suffice to give merely a broad outline of what the innovation can do, without fully disclosing the method.

Rethinking Inventorship and Suggestions

The necessary reforms required in the field, as per global suggestions and standards could be:

1. Introducing a legal definition of the word “Inventor.” Section 6 of the Patent Act, 1970 or Section 2(1)(s) should be modified to establish that the inventor is necessarily a person and that the applicants must state whether any AI was used to create the invention.
2. Regulating Ownership Default Rules where they should be implemented such that the rights are defined for the developer, user or owner of the system unless a contrary agreement exists.
3. Working on Rules for Patent Examination where the examination should be stricter, with a larger burden of proof on checking the novelty of the invention since AI can come up with multiple ideas in a short period of time, and require more information about sources, algorithms used, and the people involved in the process.
4. Conducting tests before implementing the changes and mandating the creation of a regulatory sandbox with WIPO for testing the innovations.
5. Implementing safety measures becomes very crucial. The system must include compulsory licensing and open innovations, to prevent the situation when patents are owned by a limited number of players.

Some experts suggest creating a sui-generis regime for works created by AI programs. However, it would require legislative adjustment and may contradict the TRIPS agreement and the legislation of those countries that do not recognize AI as an inventor.

Author: Akshi Seem, Associate Partner and Divya Singh and Spandan Mohapatra, Intern