By Lucy Rana and Pranit Biswas
The human-authorship starting point
Copyright in India protects original works of authorship, and the scheme of the Copyright Act, 1957 assumes an author who is a person. For literary, dramatic, musical and artistic works, authorship vests in the human creator; for a computer-generated work, Section 2(d)(vi) identifies the author as “the person who causes the work to be created.” The provision was drafted long before generative AI, but its orientation is clear: the law looks for a human at the origin of the work, not a machine. Although one has to keep in mind that the law was originally drafted in 1957 and although there have been numerous amendments over the years, the question of non-human authorship is one which does merit further review as the usage of generate AI for creation of works has increased exponentially in the last few years and the same will keep on increasing.
This creates a threshold question for any output produced with a generative system. Where a human exercises meaningful creative choice — selecting, arranging, directing, curating and materially shaping the result — there is a candidate human author, and the output may be protectable as their work. Where the output is produced autonomously by the system with little more than a short prompt, it is far less clear that any protectable authorship exists at all. The spectrum of human involvement, rather than the mere fact that AI was used, tends to drive the analysis.
The question is rarely “did AI make this?” It is “was there enough human authorship to found a copyright — and if so, whose?”
Can the machine be the author or inventor?
On the current position in India vis-à-vis the Copyright Act, 1957, the answer is an emphatic NO. The Computer Related Inventions guidelines issued by the Indian Patent Office confirm that an AI system cannot be named as an inventor on a patent application; the inventor must be a natural person. The copyright side proceeds from the same human-centred premise. An episode frequently cited in this debate is the treatment of an AI-assisted artwork whose copyright application named an AI tool as a co-author, which the Copyright Office is understood to have questioned — a reminder that Indian practice has not accepted non-human authorship. In this regard, reference is drawn to the artwork titled Suryast using generated via an AI tool called “RAGHAV” by Mr. Ankit Sahni a few years ago.
The Government has, to date, taken the view that the existing framework can accommodate AI-related works and innovations. In February 2024, the Ministry responsible told Parliament that the current patent and copyright framework was adequate and that separate amendments were not then considered necessary. At the same time, the Department for Promotion of Industry and Internal Trade constituted an expert committee in 2025 to examine whether the Copyright Act adequately addresses generative AI — an acknowledgement that the questions are live even if the statute has not changed.
Why this matters for businesses building on AI output
For a startup shipping AI-assisted code, a studio generating visual assets, or a company producing marketing text through a model, the gap between “the output exists” and “someone holds enforceable rights in it” has real consequences. Three themes recur in how businesses approach the issue:
- Protectability of the product. If material generated with little human input may not attract copyright, a business cannot assume it holds an exclusive, enforceable right in that material against copyists — which affects how defensible a product built largely on raw AI output really is.
- Contractual allocation. Because the statutory position is uncertain, the terms of the generative platform, the commissioning or vendor agreement, and employment arrangements often determine who may use and control the output as between the parties, whatever the underlying copyright position.
- Evidence of human authorship. Where protection is sought, the human creative contribution — selection, arrangement, iteration and editing — is what the analysis looks to, and records of that contribution become relevant. That being said, may not be straightforward to prove that the output was largely AI generated, with very little human intervention. Although of late, AI tools are embedding watermarks in works generated using the tools.
None of this is settled, and the committee process may in time reshape it. What is clear is that the human-authorship premise remains the organising idea of Indian copyright, and that businesses relying on generative output are, for now, operating in an area where ownership cannot simply be assumed.
| ABOUT THE FIRM
S.S. Rana & Co. is a full-service intellectual property and commercial law firm established in 1989, with offices in New Delhi, Mumbai, Chennai, Bengaluru and Hyderabad. Its practice areas include copyright, patents, trademarks, designs, technology and media law, and intellectual property dispute resolution. For further information on AI & Copyright law write to us at info@ssrana.com |