In short: Under AI generated content copyright India law, there is no clear answer yet. The Copyright Act, 1957 requires an author to be a human or legal entity — AI qualifies as neither. Whether you can claim copyright over AI-assisted content depends on how much creative skill and judgment you actually contributed.
Key points
- The Copyright Act, 1957 does not mention artificial intelligence anywhere. Its definition of “author” implicitly requires a human or recognised legal person, so AI itself cannot own copyright.
- Section 2(d)(vi) of the Act defines the author of a computer-generated work as “the person who causes the work to be created” — but that language was written in 1994 and does not account for today’s generative AI tools, which can produce outputs with minimal human direction.
- The Supreme Court’s ruling in Eastern Book Company v. D.B. Modak sets the originality standard: content must reflect adequate skill and judgment from a human. Purely AI-generated output, with little or no human creative input, is likely to fail this test.
- India has no regulations yet on using copyrighted works to train AI models — nothing comparable to opt-out or transparency rules seen in other jurisdictions.
- The Delhi High Court is currently examining a landmark case (ANI v. OpenAI) that may shape how Indian courts treat AI training data and copyright infringement.
- A copyright registration briefly recognising AI as a co-author was later revoked, underlining just how unsettled this area of law remains.
What does the Copyright Act, 1957 actually say about AI?
The short answer: nothing. The Act was last substantially amended on the question of computer-generated works in 1994, and the word “artificial intelligence” does not appear in it at all.
Copyright law in India is built on the idea that authorship vests in a natural person or a recognised legal entity such as a company. AI is treated by the law as a tool — like a camera or a word processor — not as a creator in its own right.
The “person who causes the work to be created” test
Section 2(d)(vi) of the Copyright Act is the provision that comes closest to covering AI-generated content. It defines the author of a computer-generated literary, dramatic, musical, or artistic work as “the person who causes the work to be created.”
When Parliament inserted that phrase in 1994, computer-generated work still required significant human programming and direction. A human was clearly “causing” the output. Today, a person can type a short prompt into a generative AI tool and receive a full article, image, or piece of music with very little further involvement. Whether that person has genuinely “caused” the work in any meaningful legal sense is deeply uncertain.
Courts have not yet definitively answered this. The honest position is that the provision creates more questions than it resolves when applied to modern generative AI.
When can you claim copyright over AI-assisted content?
Indian law does not bar you from owning copyright in content that AI helped you create — it depends on your level of creative contribution. The key question a court would ask is whether a human being exercised adequate skill and judgment in producing the final work.
The Supreme Court’s standard from Eastern Book Company v. D.B. Modak is instructive here. The Court rejected both the “sweat of the brow” approach (effort alone is enough) and a high bar requiring the work to be novel and non-obvious. Instead, it adopted a middle path requiring “adequate skill and judgment.” Mechanical or routine effort does not qualify; genuine creative choices do.
Applied to AI content, this suggests the following general framework:
| Scenario | Level of human creative input | Likely copyright outcome |
|---|---|---|
| You write a detailed, creative prompt; heavily edit, restructure, and curate the AI output | High | Copyright likely vests in you as the human author |
| You write a moderately specific prompt; make some editorial choices over the output | Medium | Uncertain — courts would examine facts closely |
| You type a generic prompt; publish the AI output as-is with no changes | Minimal or none | Copyright protection unlikely; work may fall into public domain |
This will be assessed case by case. There is no bright-line rule yet under Indian law. If protecting your AI-assisted work matters to your business, document your creative process carefully — your drafts, your prompt iterations, your editorial decisions.
The RAGHAV case: a cautionary tale
India briefly appeared to move toward recognising AI as a creative contributor. In a copyright registration involving an AI tool named RAGHAV, the Copyright Office initially registered a work listing the AI as a co-author. That registration was subsequently revoked.
The episode illustrates two things: first, that the Copyright Office itself is grappling with how to handle these applications; and second, that the law as it currently stands does not support AI authorship — even if an applicant can get a registration issued, it may not survive legal scrutiny.
ANI v. OpenAI: the case to watch
In November 2024, Asian News International filed a copyright infringement lawsuit against OpenAI in the Delhi High Court. ANI alleged that ChatGPT was trained on its copyrighted news articles — including paywalled content — without permission, and that the tool sometimes reproduces ANI’s reporting verbatim or attributes fabricated content to ANI.
The suit seeks an injunction and removal of ANI’s content from OpenAI’s training datasets. In its order of 19 November 2024, the Delhi High Court framed four key questions, including whether storing copyrighted data for AI training amounts to infringement and related issues of reproduction and fair use. The case is ongoing.
This is the first major Indian litigation to directly confront the AI-training-data question, and the Delhi High Court’s eventual findings will be closely watched by every founder and creator working with AI tools.
What about using others’ content to train your AI?
India currently has no specific law on this. There are no provisions requiring consent from rights-holders before using their works in an AI training dataset, no opt-out mechanisms, and no transparency obligations of the kind that exist in some other jurisdictions. This is a significant regulatory gap that Parliament has not yet addressed.
That does not mean using copyrighted content for training is automatically permitted. General copyright infringement principles under the 1957 Act still apply, and the ANI case may clarify how courts will read those principles in the AI-training context.
If you are building an AI product in India, this is an area where you should take legal advice sooner rather than later. You can also explore our broader guides on intellectual property and business law at Law for You on The Courtroom, which cover related topics in plain language.
Practical steps for founders and creators
Given the uncertainty, here is what you can do now to protect your position.
First, document your creative process. Keep records of your prompts, drafts, revisions, and editorial decisions. If your copyright is ever challenged, this evidence of human creative input is what you will rely on.
Second, do not simply publish raw AI output and assume it is protected. The less human creativity involved, the weaker your claim. Use AI as a tool within a human-led creative process, not as a replacement for it.
Third, review the terms of service of the AI platform you use. Some platforms assert rights over outputs; others assign them to you. This affects what you can legally do with the content.
Fourth, stay alert to developments in the ANI v. OpenAI case and any legislative proposals. This area of law is moving, and a ruling or new statute could change your obligations quickly.
Frequently asked questions
Can AI itself be listed as the author of a copyrighted work in India?
No. Under the Copyright Act, 1957, only a human being or a recognised legal entity such as a company can be an author. AI is treated as a tool, not a legal person. A registration that listed an AI as co-author in India was subsequently revoked, confirming that the law does not support AI authorship at present.
If I write a detailed prompt and edit the AI’s output, do I own the copyright?
You are in a stronger position if you exercise genuine creative skill and judgment — in crafting your prompt, selecting, editing, and reshaping the output. Indian courts have not yet ruled definitively on this, but the originality standard from Eastern Book Company v. D.B. Modak suggests that meaningful human creative contribution is what protects your work. Document your process carefully.
Is it legal in India to use copyrighted content to train an AI model?
There is no specific Indian law addressing this yet. General copyright principles under the Copyright Act, 1957 still apply, and using copyrighted works without permission could amount to infringement. The Delhi High Court is examining exactly this question in the ongoing ANI v. OpenAI case. Until there is a ruling or new legislation, anyone building AI products on third-party content should seek specific legal advice.
This article is for general information only and is not legal advice. Laws change; verify against the primary sources cited and consult a qualified advocate for your situation.
