Introduction

Artificial Intelligence (AI) has revolutionized innovation across sectors — from healthcare and education to media, biotechnology, and autonomous technologies. AI systems today can analyze data, design products, and even create new inventions without human intervention. This progress has raised critical legal and ethical questions: Can an AI system be an inventor? Can inventions autonomously generated by AI be patented?

In India, while the Patents Act, 1970 provides a robust framework for protecting human innovation, it does not yet explicitly address the challenges posed by AI-generated inventions. To understand the current legal position, one must examine both the letter of the law and emerging global trends shaping this evolving debate.

1. Legal Framework under the Indian Patents Act, 1970

The Patents Act, 1970 governs the grant, protection, and enforcement of patents in India. However, it was enacted at a time when AI was not envisaged as an independent creator or inventor. As such, the Act does not expressly define “inventor” or “invention” in the context of AI.

Section 6 – Persons Entitled to Apply for Patents

Under Section 6(1), a patent application may be made by: The true and first inventor of the invention, The assignee of the true and first inventor, or The legal representative of a deceased true and first inventor.

The Act assumes that the “true and first inventor” is a natural person — someone capable of conceiving an idea, assigning rights, and bearing legal obligations. This interpretation excludes AI systems, which currently lack legal personhood.

Section 2(1)(y) – True and First Inventor

While this provision does not define the term in positive terms, it specifies that the true and first inventor: “Does not include the first importer of an invention into India or a person to whom the invention is first communicated from outside India.” This phrasing reinforces the idea that an inventor must be a human originator, not an inanimate or artificial entity.

2. Patentability of AI Systems Themselves

When considering whether AI systems — such as neural networks, algorithms, or machine learning models — are patentable, one must analyze whether they satisfy the essential criteria of patentable subject matter under Indian law.

Section 3(k) – Exclusion of Computer Programs

Section 3(k) of the Patents Act, 1970 states that the following are not inventions:

“A mathematical or business method or a computer program per se or algorithms.”

This provision directly impacts AI-related inventions, as most AI systems are software-driven and involve mathematical algorithms. Consequently, AI systems as such are not patentable unless they demonstrate a technical contribution beyond mere computer programming.

Patent Office Guidelines

The Indian Patent Office (IPO) issued Guidelines for Examination of Computer Related Inventions (CRIs) in 2017, which clarify that: A computer program per se is not patentable.

However, if the invention demonstrates a technical effect or technical advancement, such as improved hardware performance or a novel industrial application, it may be considered patentable.

Therefore, while the AI algorithm itself may not be patentable, an AI-based system integrated with hardware (e.g., a self-driving mechanism or medical diagnostic tool) that produces a tangible technical result could qualify for patent protection.

3. AI-Generated Inventions and Inventorship

The more complex question concerns inventions created autonomously by AI systems without direct human input. Who should be listed as the inventor in such cases — the AI system, its programmer, or its owner?

Current Legal Position in India

Under current Indian law: Only a natural person can be listed as an inventor in Form 1 (Patent Application). AI lacks legal personality, and therefore cannot: Hold ownership rights, Assign or transfer rights, Take an oath or declaration as required under patent formalities.

Thus, AI cannot be recognized as an inventor. However, if a human researcher designs, trains, or supervises the AI system that produces an invention, that human can be regarded as the true and first inventor.

4. Global Perspective: The DABUS Case

The question of AI inventorship came to the forefront globally with the DABUS case (Device for the Autonomous Bootstrapping of Unified Sentience), developed by Dr. Stephen Thaler.

Facts of the Case

Dr. Thaler filed patent applications worldwide naming DABUS — an AI system — as the inventor. The applications were filed in the U.S., U.K., E.U., Australia, and other jurisdictions.

Judicial and Administrative Outcomes

Jurisdiction	Decision	Legal Reasoning
United States (USPTO)	❌ Rejected	Inventor must be a natural person under the U.S. Patent Act.
United Kingdom (IPO & Court of Appeal)	❌ Rejected	AI lacks legal personality.
European Patent Office (EPO)	❌ Rejected	Only a natural person can be named as inventor.
Australia (Federal Court)	❌ Reversed earlier acceptance	AI cannot own or transfer patent rights.
South Africa (CIPC)	✅ Accepted (administratively)	Application accepted without substantive examination; not a precedent.

India has not yet faced a similar case, but given the statutory framework and IPO practice, it is highly likely that AI would not be accepted as an inventor here either.

5. Patentability of AI-Assisted Inventions

While AI cannot be listed as an inventor, inventions developed with the assistance of AI can be patentable if certain conditions are met. The human inventor who directs, controls, or interprets the AI’s output can claim inventorship.

Criteria for Patentability

For an AI-assisted invention to be patentable under Indian law, it must satisfy the following:

Novelty – The invention must be new and not part of prior art. Inventive Step – It must not be obvious to a person skilled in the art. Industrial Applicability – It must be capable of industrial application. Human Inventorship – A human must be identified as the “true and first inventor.” Technical Contribution – The AI must contribute to solving a technical problem, not merely perform computations.

6. Ethical and Policy Challenges

The rise of AI inventorship raises significant ethical and policy questions:

Should AI be granted legal personhood or partial recognition for inventorship?

How can the law balance human creativity with machine-generated innovation?

If AI cannot own IP rights, should ownership vest in its developer, user, or owner?

Many legal scholars advocate for a “human-centric” approach, where AI is treated as a tool, and its human controller remains responsible for inventive acts. Others argue that, as AI systems evolve, future amendments may be necessary to recognize AI contribution in patent documentation.

7. The Way Forward for India

India, being a technology-driven economy, must proactively address the intersection of AI and intellectual property. The following steps could shape future policy:

Issuing guidelines on AI-related patent applications by the Indian Patent Office. Amending the Patents Act to clarify the definition of “inventor” and “invention” in the context of AI. Encouraging AI-human collaboration models where human oversight remains essential. Promoting ethical AI frameworks to ensure transparency and accountability.

Conclusion

Under the current Indian Patents Act, 1970, AI systems themselves are not patentable, and AI cannot be named as an inventor, as the law recognizes only natural persons. However, inventions created with the assistance of AI may qualify for patent protection if a human inventor is identified and the invention meets the established criteria of novelty, inventive step, and industrial applicability.

While AI continues to transform the landscape of innovation, Indian law remains grounded in a human-centric philosophy of intellectual property. As technology advances, India may eventually need to revisit its legal definitions to accommodate AI’s growing role in creation and invention — ensuring that innovation continues to thrive within a clear, equitable, and forward-looking legal framework.