Supreme Court's Amicus Review In The Human-AI Integration Patent Dispute

Introduction : In a special leave petition filed regarding the rejection of a patent application for an invention which its inventor claims to be a process for improving human capabilities by integrating with artificial intelligence, Senior Advocate Swathi Sukumar was appointed as the amicus curiae in the Supreme Court of India on July 21, 2026 (Caleb Suresh Motupalli v. Controller of Patents).  The application, which became known to the Controller of Patents as a device and a method of super-augmenting a persona to manifest what the inventor termed a Pan-Environment Super-Cyborg, had been refused by the Controller of Patents and on appeal by the Madras High Court. The bench of Justice K.V. Viswanathan and Justice Alok Aradhe directed the amicus to provide technical and legal assistance to the Court and the matter is further adjourned for hearing on August 24, 2026.

The problem goes beyond the odd facts of the case. It forces the Court in this case, possibly for the first time at the highest level, to confront, if it can confront, the application of the patentability tests in the Patents Act, 1970 to inventions that claim to fuse the cognitive or physical capabilities of human beings with machine intelligence. In this blog, we shall discuss the patentability issues the case raises, the need for the appointment of the amicus and the lessons for applicants seeking to draft AI-driven inventions in India.

Legal Provisions

A. The Statutory Framework

Patents Act, 1970, defines patentable invention as a new product or process having an inventive step and capable of being applied to industry. According to Section 2(1)(ja), inventive step exists when a feature is considered to offer technical advance over the known state of the art, or it is economically significant, or it has both these characteristics and one would not be obvious to a person skilled in the art. In order to enable a person skilled in the art to practise a person according to the invention without undue experimentation, the complete specification must be able to describe and specify the invention and its operation completely and particularly and disclose the best method of practising it known to the applicant at the time of filing the specification. 

Section 3(k) excludes from the definition of invention a mathematical or business method, a computer program per se, or an algorithm; and Section 3(m) excludes a mere scheme, rule or method of performing a mental act. Section 59 limits amendments to claims during prosecution to those that are within the scope of the specification as originally filed in the application, and Section 117-A prescribes the types of amendments that can be made in a statutory appeal.

B. The Motupalli Prosecution History

Indian National phase Application was filed by Motupalli in June 2012. After several examinations, the Controller of Patents in April 2021 refused the application due to various reasons, including the absence of enablement and clarity, the claims having been amended beyond the scope of the original disclosures, the absence of any inventive step, and non-patentability under the Act. In October 2021 another review request by the applicant was turned down.

On appeal from the said order passed in review under Section 77(1) (f), the Madras High Court, held that no appeal is maintainable against an order passed in review under Section 77(1) (f) of the Scheme, and that the same is not a scheme of review under Section 117-A of the Scheme as stipulated in the proviso to Section 116 of the Scheme. In considering the matter on merits without prejudice to that finding, the Court recognized that the complete specification was insufficient to enable a person of skill in the art to practice the invention and that the application had aggregated prior art of both AI and human augmentation without providing any technical details on how the claimed “persona-extension effect” was achieved.

Legal Analysis

A. Inventive Step and the Aggregation Problem

The violation of the Inventive Step requirement and the Aggregation Problem. One common challenge with claims involving AI integration is that of identifying a true composition of new elements rather than an assembly of known ones. When a combination of a known type of human augmentation hardware and a known type of AI-based processing is used in an application, and the combination does not show a technical synergistic effect, then the combination may be considered obvious under 2(1)(ja). This is exactly the concern that led the Madras High Court to conclude that the specification had merely brought together what was already disclosed. The question of “inventive step” is not whether components are known prior to the specification, but whether they have an effect together that wasn't obvious to a person skilled in the art. In human-AI integration inventions, this demands that the applicants define and demonstrate how the human and AI components have synergistic features, which could be in latency, signal fidelity, accuracy of the sensor fusion etc., and not just on the novelty of the idea itself.

B. Enablement and the Problem of Result-Oriented Claiming

For AI inventions, especially if targeting a desired result instead of a specific technical method to accomplish the result, section 10(4) is often the more significant limitor. An enhancement in one or more of cognition, persona or enhanced human capability requires a mechanism, i.e., a description of the architecture, data flow, algorithmic steps, or hardware/software interfaces that produce that enhancement, not just a result; the specification must provide such a description, in sufficient detail so that it can be replicated without undue experimentation. The Madras High Court observed that the specification did not disclose the technical bridge between the input and output is a typical enablement failure, something was claimed but the technical steps that would make it possible to achieve the claimed augmentation were not disclosed. This is a common issue with AI related filings in general, since the underlying model could be a black box even for the AI that is filing the patent.

C. The Boundary Between Human Contribution and Machine Functionality

The novel question in the case is probably where a human-AI integration invention falls in relation to the Section 3(k) computer program exclusion and the Section 3(m) mental act exclusion. Even if there is some AI hardware involved, an invention that primarily reorganises or supplements human cognitive or perceptual processes is likely to be deemed a scheme for mental acts instead of a technical solution to a technical problem. On the other hand, if the invention is framed in terms of the human interface, but incorporates AI to directly improve a physical device, like a sensor processor, data transmission, or physical control, it has a better chance of being patentable. This boundary is not new to Indian patent law, but claims for integration of AI further hone the boundary as the work of the human and the machine are described in the specification as one, and they are not technically separated. Notably, the Delhi High Court has in parallel issued directions to establish guidelines for differentiating mental acts from patentable inventions in its pending case in 2026, signaling a concerted effort by Indian courts to clarify this distinction in various proceedings.

Why the Amicus Appointment Matters

The Supreme Court's decision to appoint a senior patent practitioner to be amicus curiae, specifically to help on technical and ethical issues, is noteworthy for a number of reasons. First, the underlying issue is a self-represented inventor and a truly novel category of invention; the amicus appointment reflects the Court's understanding that it will need a fair solution that requires framework beyond either party's capacity, and that will require the participation of an amicus. Second, the appeal court decision in this case will not be binding on the court of appeal, and it will be persuasive to a broader category of human-AI integration patent applications, not just this case. Third, the amicus brief, filed, is likely to serve as a guide for others to patent offices and applicants attempting to solve comparable problems.Third, once filed, the amicus brief will likely become a reference for other patent offices and applicants involved in similar problems because of the lack of settled doctrine in this area. The appointment is not a mere formality, but rather an implicit recognition that the case represents a real "frontier" in Indian Patent Law.

Drafting Lessons for AI-Enabled Inventions

• Focus on specific technical solutions (such as specific hardware components, sensor architectures, model structures or data pipelines) and not on aspirational solutions (such as augmented cognition or enhanced persona).

• Enable disclosure of the algorithmic and structural steps between the input and output that are claimed, such that a person of ordinary skill in the field could implement the invention without undue experimentation.

• Do not black-box claim AI; if AI is a key part of the invention, describe its function, training or integration logic, as necessary to enable the invention, even if the commercial embodiments are proprietary.

• Don't rub shoulders on the novelty of combining known human-augmentation and AI in the abstract, but make a clear claim on a specific measurable technical impact of the human-AI combination.

• Proactively address the Section 3(k) and Section 3(m) exclusions in the specification by separating the claimed invention from a computer program per se and/or from a scheme for performing a mental act.

• Avoid Section 59 added-matter objections by restricting amendments to claims during prosecution to the claims as originally filed.

• In this area of invention, input from technical experts and comparative data are most likely to be closely reviewed by examiners and courts during prosecution, and should be prepared in advance.

• If the broad, conceptually ambitious claim is commercially desirable, discuss a Phase Filing with a series of narrow Divisional Claims based on specific embodiments, and not a single wide claim that may be rejected.

Conclusion

The Motupalli reference will take the Supreme Court through a series of questions regarding patentability that will likely become more frequent as AI-integration inventions progress beyond the speculative realm into areas of commercial technology, including wearables, neural interfaces and assistive devices. The case is a compressed mini-tutorial of the three doctrinal "hot spots" that inventions commonly encounter: whether the claimed combination is a genuine invention or merely a collection of known elements; whether the specification describes how to make the claimed invention, but does not assert it; and whether the invention's technical contribution can be distinguished from its underlying scheme for enhancing human cognition or behavior. Regardless of the outcome, the amicus curiae brief and the Court's reasoning will likely impact patent offices and applicants in human-AI integration patent claims for years to come and careful, technically-sound drafting of these types of patent claims is even more important than ever.

Author: Prince Lucky Jain in case of any queries please contact/write back to us via email to content@khuranaandkhurana.com or at  Khurana & Khurana, Advocates and IP Attorney