AI Patentability in India, Europe, and the US: Patentability Standards and Emerging Trends

Recent advancements in Information and Communication Technology (ICT), including Artificial Intelligence (AI) are rapidly transforming industries. As AI  becomes integral to modern products, securing patents has become crucial for fostering innovations.

This article delves into the patentability of AI related inventions in India, comparing the Indian Patent Office (IPO) framework with that of the European Patent Office (EPO) and the United States Patent and Trademark Office (USPTO).

Legal Framework for AI Patentability under Indian Patent Law

In India, the patentability of AI-related inventions is guided by the 2017 CRI Guidelines and Section 3(k) of the Patents Act. While the 2017 Guidelines do not explicitly address AI, they allow patenting of software-based inventions showing technical advancement. While there are not many judicial precedents, a recent ruling by the Hon’ble Madras High Court in Caleb Suresh Motupalli v. Controller of Patents (2025) provides some judicial guidance on assessing AI inventions.

Additionally, the recently published Draft CRI Guidelines 2025 propose a more structured approach for evaluating AI related invention, clarifying that AI-assisted inventions are patentable provided they demonstrate technical effect. The guidelines further stress on the nature of disclosure of specification highlighting that the training data should be adequately defined, requiring a link between the data’s characteristics and the technical problem being addressed—aligning closely with the EPO’s approach.

Key Judicial Precedents

1. Caleb Suresh Motupalli vs Controller of Patents (2025):  In this case, the Court identified key patentability requirements under Indian law, especially for AI-related inventions. It clarified that AI-related inventions must disclose how the claimed invention works in sufficient detail to allow others to reproduce it without undue experimentation. The Court stressed that merely combining existing AI techniques without providing clear, technical instructions or working examples does not satisfy the enablement requirement.

   
   

2. Ferrid Allani v. Union of India (2019): The Delhi High Court clarified that inventions demonstrating a technical effect and technical contribution are patentable, even if based on computer programs. The Court acknowledged that AI and blockchain technologies should not be excluded merely for being software-based.

   
   

3. Microsoft Technology Licensing, LLC v. Assistant Controller of Patents and Designs (2023): The Court emphasised that invention should not be deemed a computer program per se merely because it involves algorithms and computer-executable instructions; rather, it should be assessed based on the technical advancements it offers and its practical application in solving real-world problems.

   
   

Patentability of AI-Related Inventions in India, with insights on EPO and USPTO practices 

The key to patenting an AI invention in India lies in demonstrating a technical contribution or effect, which largely follows the European approach, i.e. focusing on technical effects and contributions.

While EPO has provided specific guidelines for assessing AI innovations, there are no specific guidelines for AI related inventions in India yet. EPO has further clarified that if technical effect relies on specific properties of the training data, those properties must be described. While there is no need to disclose the exact training dataset, sufficient information about how to generate suitable training data should be provided as per the EPO guidelines for AI related Innovations.

USPTO’s 2024 guidance update clarifies AI-assisted inventions are generally patent-eligible, in line with the Alice/Mayo framework, with an emphasis on the integration of abstract ideas into practical applications.

In India the Draft CRI Guidelines 2025 propose to include guidance on examination of AI related inventions, and we expect to see refinement of patent laws and guidelines for examination of AI related inventions in the final CRI guidelines. 

Comparative Analysis: India, USA, and EP

Aspect	India	USA	EP
Rules & Guidelines	Section 3(k), CRI Guidelines (2017)	Alice/Mayo two-step test for subject matter eligibility	Article 52(2), EPO Guidelines (G-II, 3.3)
Exclusions	Mathematical methods, business methods, computer programs per se, algorithms	Abstract ideas, laws of nature, mathematical methods	Methods for mental acts, business, games, computer programs per se
AI invention Patentability	Allowed if technical effect, and improvement in system functionality	Allowed if offering practical application and technical implementation	Allowed if showing technical effect
General Approach	Focus on technical effect & contribution, and sufficiency of disclosure	Focus on practical application	Focus on technical effect & contribution, and sufficiency of disclosure

Key Insights

• USPTO: Broader approach, focusing on practical applicability.

• EPO: Requires technical contribution and technical character. If invention relies on specific properties of the training data, those properties must be sufficiently described. No specific need to disclose the exact training dataset.

• IPO: Similar to EPO, requires technical contribution and sufficiently disclosed. Invention should have a technical, tangible effect.

  
  

AI Patent Trends in India

Growth in Filings

India has witnessed a significant rise in AI-related patent filings over the past few years. This surge reflects the growing emphasis on Artificial Intelligence (AI) across various industries.

Domestic vs. Foreign Applicants

Initially, the majority of AI patent applications in India were filed by foreign applicants with technology giants leading the way. However, with government initiatives (such as Digital India and AI for All) there has been a steady increase in domestic filings as Indian companies, startups, and research institutions ramp up AI-related innovations.

Sample Granted AI-Related Patents:

Here are a few examples of AI-related patent applications granted by the IPO.

1. AI-Based Fruit Ripening Detection (Patent No. 561369): A non-invasive method using AI to detect artificial fruit ripening.

2. AI & IoT-Based Water Hyacinth Management (Patent No. 555713): An automated system for real-time detection and removal of water hyacinth.

3. AI-Based Traffic Estimation (Patent No. 562065): Automatic and accurate traffic count estimation.

4. AI-Based Speech-Assisted Automation System (Patent No. 430666): A multi-language speech recognition system for computer automation.

5. AI-Enabled Targeted Content System (Patent No. 555866): Combines deep learning with blockchain for content delivery.

6. AI Processor for Neural Network Operations (Patent No. 555670): Reduces hardware resource and power consumption during image processing.

Conclusion

The patentability of AI-based inventions in India, Europe, and the United States is governed by distinct standards, although certain common principles apply across these jurisdictions. A fundamental requirement in all three systems is that the claimed invention must go beyond a mere abstract idea.

In the Europe as well as in India, a key consideration is whether the invention addresses a specific technical problem and provides a technical solution. Both jurisdictions emphasize the need for a technical effect that contributes to the functioning of a technical system or process. The U.S. approach on the other hand is more rooted in judicial precedents focused on abstract ideas, placing less emphasis on demonstrating a technical effect.