Imagine a researcher using artificial intelligence to identify a new material, a pharmaceutical team using machine learning to discover a promising compound, or an engineer using an AI system to develop a more efficient industrial process. The technology may have helped uncover the solution. But who is the inventor? And can the resulting invention receive a patent in India?
These questions are becoming increasingly important as AI moves from assisting research to shaping the inventive process itself. Indian patent law does not prohibit every invention developed with AI. The more difficult questions concern human inventorship, technical contribution and whether the claimed invention satisfies the statutory requirements for patent protection.
When AI Becomes Part of the Inventive Process
Under Section 2(1)(j) of the Patents Act, 1970, an invention must be a new product or process involving an inventive step and capable of industrial application. It must also avoid the exclusions under Sections 3 and 4.
Section 3(k) excludes a mathematical or business method, a computer programme per se and algorithms. This creates a particular challenge for AI-related inventions. A claim directed merely to an abstract algorithm or mathematical model may fall within the exclusion. An invention involving AI, however, is not automatically excluded simply because software or an algorithm forms part of it.
The distinction lies in what the claims actually seek to protect and whether the invention demonstrates a qualifying technical contribution.
What the 2025 Guidelines Clarify
In July 2025, the Indian Patent Office released its revised Guidelines for Examination of Computer Related Inventions (CRIs). The guidelines include a dedicated discussion of AI, machine learning and deep learning, with examples addressing patentable and non-patentable subject matter.
They distinguish between AI-generated inventions, developed autonomously or with very limited human intervention, and AI-assisted inventions, in which AI serves as a tool in the inventive process.
The guidelines state that AI-generated inventions cannot identify an AI system as the true and first inventor under the current statutory framework. AI-assisted inventions, by contrast, are not categorically excluded under Section 3(k), provided they meet the applicable patentability requirements and demonstrate the necessary technical effect through their inventive application.
The distinction is significant. Using AI does not, by itself, disqualify an invention. Equally, describing an AI model as innovative does not establish that its underlying invention is patentable.
The guidelines are examination guidance, not a substitute for the Patents Act or judicial decisions. The statutory requirements continue to determine the outcome.
What Indian Courts Have Said About AI-Related Inventions
In Ferid Allani v. Union of India, the Delhi High Court examined the patentability of a computer-related invention under Section 3(k). The Court emphasised that computer-related inventions should not be rejected merely because they involve computer programmes. Where an invention demonstrates a technical effect or technical contribution, its patentability requires substantive examination.
Although the decision predates the current AI boom, its reasoning remains relevant. An AI-related patent application must be assessed by examining the claimed invention and its technical contribution, rather than treating the presence of software as an automatic bar.
A more recent illustration of the challenges involved in computer-related patent applications is Caleb Suresh Motupalli v. Controller of Patents, decided by the Madras High Court on 29 January 2025. The application concerned a proposed “super-augmented persona” system. The Court examined the claimed invention and the Patent Office’s objections, including issues concerning the clarity and sufficiency of the disclosure.
The case highlights a separate but important principle: an inventive concept must be explained with sufficient detail to enable the invention to be understood and implemented. Broad descriptions of what a sophisticated system might achieve cannot necessarily substitute for a clear technical disclosure.
Neither case establishes that all AI-assisted inventions are patentable. Together, however, they illustrate why technical substance and the quality of the patent specification matter.
Who Should Be Named as the Inventor?
Patent law also requires a distinction between inventorship and ownership. Under Section 6, a patent application may be made by the true and first inventor, an assignee, or a legal representative in the circumstances specified by the Act.
A company may own patent rights through a valid assignment, but that does not make the company the human inventor. Likewise, using an AI tool does not automatically make its developer, operator or user the inventor.
Where AI contributes to an invention, the human contribution must be assessed carefully. The question is not simply who operated the software. It is who qualifies as the inventor of the claimed subject matter under the applicable law. Indian law does not currently recognise an AI system itself as a patent inventor.
The Question Is Not Whether AI Was Used
AI-assisted invention presents Indian patent law with a question more precise than whether machines can innovate. It asks whether the claimed invention meets the legal requirements for protection and whether a legally recognised inventor can be identified.
For businesses investing in AI-driven research, that distinction has practical consequences. Technical records, the contribution of researchers, the definition of the invention and the quality of disclosure may all matter when a patent application is examined or challenged.
AI may change how inventions are discovered. It does not remove the need to establish what the invention is, why it qualifies for protection and who legally invented it.
etc. whilst wrongfully claiming to be part of our firm and making false claims and allegations.