Does the Indian Patent Office Use AI? Full Guide

Artificial intelligence is increasingly being used by patent offices globally for prior-art searching, classification and administrative processing. This raises an important question for applicants considering patent registration in India: does the Indian Patent Office (IPO) use AI in patent examination in India, and what legal limits would apply?

As of July 2026, the IPO has not publicly identified a named, examiner-facing AI system comparable to tools disclosed by certain overseas patent offices. This does not establish that no automation or AI is used within the Office. It means that there is presently no published framework explaining whether AI contributes to examination, how its output is validated or how responsibility is divided between technology and the statutory decision-maker.

The distinction matters. AI may assist an examiner, but examination and decision-making under the Patents Act, 1970 remain entrusted to designated human authorities.

What is publicly known?

The IPO has substantially digitised patent administration. Applications are filed and processed electronically; patent records are searchable online; prosecution documents are available through the electronic register; hearings are conducted by video conference; and applications are allocated through centralised systems. These tools support patent registration in India and public patent search, but digitisation is not synonymous with AI.

The IPO has not publicly described its InPASS search facility as an AI-based system or as a patent search AI tool. Nor does any publicly available IPO notice, guideline or annual report presently describe a named tool used by examiners to identify or rank prior art, assess novelty or inventive step, draft examination objections or recommend grant or refusal.

What Indian Patent Attorneys Are Observing

At the same time, experienced Indian patent practitioners have observed a discernible improvement in the depth and precision of at least some examination reports and orders. In certain cases, the speed with which highly relevant prior art is identified, the claim-by-claim structure of objections and the quality of technical analysis may naturally prompt practitioners to wonder whether examiners are benefiting from advanced search or drafting assistance.

That observation should not be converted into a factual assertion. It is anecdotal, examination quality remains uneven, and several other developments may explain the improvement. The IPO’s Annual Report for 2024–25 records that 407 newly recruited patent examiners joined the Office in January 2025 and completed an induction programme. Better training, centralised allocation, greater experience and improved access to technical literature may all be contributing. The fair conclusion is therefore that technological assistance is a reasonable possibility, but no examiner-facing AI programme has been publicly confirmed.

Can the IPO legally use AI?

Neither the Patents Act, 1970 nor the Patents Rules, 2003 expressly authorises or prohibits AI as an internal examination aid. In principle, the IPO could use technology for classification, prior-art searching, translation, formality review or document management without an amendment to the Act.

There is, however, an important limit: an administrative tool cannot replace the statutory authority entrusted with examination and decision-making.

Under Sections 12 and 13, an examiner investigates compliance with the Act and Rules, including anticipation by previous publication and prior claiming, and reports to the Controller. The Controller communicates objections, considers the applicant’s response, provides a hearing where required and ultimately decides whether the application should proceed to grant, amendment or refusal.

These functions involve applying law to facts and exercising judgment. An AI system may locate documents, organise information or prepare preliminary analysis, but it cannot relieve the examiner or Controller of the obligation to assess the application independently

Essential safeguards

If AI-assisted examination is introduced—or is already being used internally—five safeguards are particularly important:

Human oversight: Every objection and decision should remain attributable to an examiner or Controller who has independently reviewed the claims, prior art and legal requirements.
Disclosure: An applicant must understand and answer the case against it. If an AI tool identifies prior art, the examination report must still identify the relevant disclosure, relate it to the claims and explain the legal objection. A high similarity score does not establish anticipation or obviousness.
Reasoned decisions: A refusal must reflect the Controller’s reasoning and address the applicant’s material submissions. Generic or unexplained conclusions would remain vulnerable to challenge under principles of natural justice.
Confidentiality: Applications are ordinarily confidential before publication under Section 11A. If an external AI system processes unpublished specifications, safeguards would be required concerning access, storage, model training, retention, security and cross-border transfers.
Testing and accountability: AI search may produce false positives or overlook relevant documents. Any system should be tested across technologies and languages, periodically audited and subject to a process through which applicants may raise concerns.

Implications for applicants

Applicants should not assume that the IPO is conducting an AI-enhanced search capable of locating every relevant document. A professionally designed prior-art search remains important for patent registration in India, particularly in crowded technical fields where AI patent search tools may assist but should not replace expert legal and technical review.

Examination reports should also be reviewed critically. Textual or semantic similarity does not itself establish anticipation. An inventive-step objection likewise requires a reasoned analysis and cannot be sustained merely by listing technologically related references.

The road ahead

AI could help the IPO manage increasing filing volumes, search multilingual sources and reduce repetitive work. It could also be one among several factors contributing to the improving quality visible in parts of Indian patent examination. Used responsibly, it would complement rather than diminish the expertise of examiners and Controllers.

The present issue is therefore not whether the profession suspects that AI may be playing some role. It is that no public framework enables stakeholders to know whether it is being used, for what purposes and subject to what safeguards. If the IPO adopts AI-assisted examination, a transparent policy addressing human oversight, confidentiality, validation and accountability would benefit both the Office and applicants.

Until then, the legally sound position remains that no examiner-facing IPO AI programme has been publicly disclosed, and statutory responsibility continues to rest with human examiners and Controllers.