{"id":20643,"date":"2026-08-14T03:09:46","date_gmt":"2026-08-13T21:39:46","guid":{"rendered":"https:\/\/lawjurist.com\/?p=20643"},"modified":"2026-08-14T03:14:42","modified_gmt":"2026-08-13T21:44:42","slug":"who-owns-ai-generated-content-in-india","status":"publish","type":"post","link":"https:\/\/lawjurist.com\/index.php\/2026\/08\/14\/who-owns-ai-generated-content-in-india\/","title":{"rendered":"Who Owns AI-Generated Content in India"},"content":{"rendered":"\t\t
\n\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\n

Author: Nikhil Kumar, \u00a0a 1st year BA.LLB(Hons.) at University of Lucknow <\/p>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t

\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t
Introduction<\/strong><\/h5>\n

Type a prompt into an AI tool today, and within seconds, you will have a poem, logo, piece of code, or full-length article sitting in front of you. Millions of people in India now use these tools every day for work and fun without pausing to ask a fairly basic question: who actually owns what comes out the other end? This may sound like a technical footnote, but it is not. If a business builds its marketing campaign around AI-generated images or a writer publishes a book drafted with the help of a chatbot, the answer to that question decides whether they can stop someone else from copying it, whether they can sell it, and whether they can even claim it as theirs in the first place. Indian law was written for a world where every creative work had a human being behind it, and it is now being stretched, tested, and argued in court to see whether it can accommodate a world where the “creator” might be a machine.<\/p>\n

The Legal Starting Point<\/u><\/strong><\/p>\n

India’s copyright framework comes from the Copyright Act, 1957, a law that has been amended several times but was never built with generative AI in mind. Section 2(d) of the Act defines who counts as an “author” for different kinds of work: a writer for literary work, a composer for music, an artist for artistic work, and so on. Buried in that same section is a clause that matters a great deal here. For any literary, dramatic, musical, or artistic work that is “computer-generated,” the Act says the author is “the person who causes the work to be created.”<\/p>\n

That phrase was written decades ago, almost certainly with something far simpler in mind, such as a spreadsheet macro or an early computer-assisted design tool, not a large language model trained on billions of documents. However, it is the closest thing Indian law has to a rule for AI output, and lawyers, courts, and the Copyright Office keep coming back to it. The wording is deliberately human-centered. It does not ask what produced the work; it asks who caused it to be produced. That framing has become the entire battlefield on which the ownership question in India is now being fought.<\/p>\n

Why the Machine Itself Cannot Own Anything<\/u><\/strong><\/p>\n

One point on which Indian law is fairly settled is that an AI system cannot hold copyright in its own name. Copyright, as a legal right, can only vest in a person, and Indian law recognizes two types of persons: natural persons, such as you and me, and legal persons, such as companies or registered trusts. An algorithm is neither. It cannot sign a licence, cannot sue for infringement, and cannot be dragged to court either. Therefore, even in cases where an AI tool appears to be doing almost all of the creative work, ownership must land somewhere else, usually with the human who set the process in motion.<\/p>\n

This aligns with how courts elsewhere have approached similar questions. In the United Kingdom, the developer behind an AI system called DABUS tried to have the machine listed as the inventor on a patent application. The UK Intellectual Property Office rejected this application, and the High Court agreed, holding that only a natural person can be named as an inventor because a machine has no legal personality. India has not seen an identical case reach its courts yet, but the underlying logic that legal rights need a legal person to attach to runs through Indian jurisprudence as firmly, and would almost certainly produce the same result here.<\/p>\n

The Sahni Case and the RAGHAV Episode<\/u><\/strong><\/p>\n

The clearest real-world test of this question in India involves an IP lawyer named Ankit Sahni and an AI painting application called RAGHAV. Sahni took a photograph he had shot himself of a sunset, then fed it into RAGHAV along with a well-known painting style, and the tool generated a new image called “Suryast,” rendered in the visual style of Van Gogh’s Starry Night. Sahni applied to register the work and, notably, tried to list RAGHAV as a co-author alongside himself.<\/p>\n

The United States Copyright Office turned down the equivalent application outright, holding firm to its position that copyright protection requires human authorship and nothing else. India’s Copyright Office initially took a different, more permissive path. It registered the work with RAGHAV named as a co-author, a decision that made headlines internationally because it appeared to break from the human-authorship principle followed almost everywhere else. However, that registration did not survive scrutiny for long. Following internal review and considerable pushback from the wider legal community, the Copyright Office moved to withdraw the registration. The episode did not produce a binding judicial precedent, since it played out largely at the registration stage rather than in a courtroom, but it sent a fairly clear signal about where Indian authorities stand: AI can be a tool a human uses to create something, but it cannot be named as a joint owner of what results.<\/p>\n

What the case supports, and has since become something like a working principle among Indian IP practitioners, is a sliding scale based on the amount of creative judgment the human actually exercised. Someone who types a single vague prompt and accepts whatever the AI produces without changes sits at one end of that scale, and their claim to authorship looks thin. Someone who directs the process closely, makes deliberate creative choices about composition, style, and editing, and reworks the AI’s raw output into something meaningfully their own sits at the other end, with a much stronger claim. Indian law has not drawn a bright line marking exactly where “enough” human input begins, and that grey zone is where most real disputes are likely to land.<\/p>\n

How Other Countries Are Handling the Same Question<\/u><\/strong><\/p>\n

It helps to look sideways at how other jurisdictions have dealt with this, partly because India’s own approach borrows pieces from more than one jurisdiction. The United Kingdom’s Copyright, Designs and Patents Act, 1988, actually goes further than Indian law in giving computer-generated works an explicit owner: Section 9(3) states that for a work generated by a computer with no human author, the author is taken to be the person who made the arrangements necessary for the work to be created. It is one of the very few legal systems in the world that grants protection to a work even when no human directly authored it, as long as someone set up the process.<\/p>\n

The United States has gone the opposite way. The US Copyright Office has repeatedly and publicly held that copyright protection is reserved for works of human authorship, and that purely AI-generated material, without meaningful human creative contribution, cannot be registered. Several applicants have tested this position in court and lost. The European Union, meanwhile, does not yet have a single unified answer, although its broader AI Act and ongoing copyright discussions are pushing member states toward more human-authorship-centered rules similar to the American approach.<\/p>\n

India is closer to the American and European positions than the British one. Because Section 2(d) still asks who “caused” the work to be created rather than granting automatic ownership over any computer output, Indian law implicitly requires an identifiable human hand in the process, even if that hand only pressed a button and picked a result. This is a meaningfully different design choice from the UK’s, and it means that Indian businesses cannot simply rely on the existence of an AI-generated asset to claim ownership; they need to be able to point to the human decisions behind it.<\/p>\n

The Bigger Fight: Training Data and the ANI versus OpenAI Case<\/u><\/strong><\/p>\n

Ownership of AI output is only one half of the puzzle. The other half, which has become the more consequential legal battle in India so far, is about what AI companies are allowed to use while training their models. This question arrived in the Delhi High Court through a case brought by Asian News International, one of India’s largest news agencies, against OpenAI.<\/p>\n

ANI alleged that OpenAI had trained ChatGPT on its copyrighted news articles without permission or a licence, and separately argued that the chatbot had, in some responses, generated false information and wrongly attributed it to ANI as if it were their reporting. The case drew unusually wide participation for an Indian copyright dispute. Publishing bodies, including the Digital News Publishers Association and the Federation of Indian Publishers, intervened in support of ANI, while industry groups such as the Broadband India Forum and several AI companies intervened on OpenAI’s side, reflecting how much the outcome mattered to both the media industry and the AI sector.<\/p>\n

On July 24, 2026, Justice Amit Bansal of the Delhi High Court delivered an interim ruling that stopped short of deciding the case but gave the clearest signal yet of how Indian courts might approach AI training. The Court declined to grant ANI the interim injunction it had asked for, holding, on a prima facie basis, that OpenAI’s storage and use of ANI’s articles to train its models could fall within the fair dealing exception under Section 52(1)(a) of the Copyright Act, the same provision that permits copyrighted material to be used for private study, research, and similar non-commercial purposes. The Court also found that ANI had not shown that ChatGPT’s responses reproduced or closely mirrored its original reporting, which weakened the infringement claim on the output side. Importantly, the ruling was explicitly interim and provisional in nature. The main suit continues, and a final judgment could still go either way once the case is fully argued. Even so, this is being treated as India’s first substantive judicial statement on how copyright law interacts with large language model training, and whatever the eventual outcome, it is likely to shape how every AI company operating in India thinks about the data it uses.<\/p>\n

What This Means in Practice<\/u><\/strong><\/p>\n

For anyone actually using these tools right now, rather than watching the litigation from a distance, a few practical points follow from where the law currently stands. First, if you want a strong claim to own something an AI helped you make, keep evidence of your own creative input: the prompts you refined, the drafts you edited, and the choices you made about what to keep and discard. A single unedited AI output, generated from a bare prompt, is the weakest position to be in if ownership is ever challenged. Second, businesses that rely heavily on AI-generated content for branding, marketing, or product design should build contracts and internal policies that clearly assign ownership of that output to the company, rather than assuming copyright law will sort it out automatically. Third, anyone using AI tools trained on scraped or licensed data should keep an eye on how the ANI dispute eventually resolves, since a different outcome at the final hearing stage could change the calculus around what training practices are considered lawful in India, with knock-on effects for pricing, access, and liability across the AI industry.<\/p>\n

Where the Law Needs to Go From Here<\/u><\/strong><\/p>\n

The honest assessment is that Indian copyright law is being asked to answer questions it was never designed for, using language written for a pre-internet, pre-AI world. Section 2(d)’s “person who causes the work to be created” has proven flexible enough to avoid an obvious legal vacuum, but flexibility is not the same as clarity. There is no statutory guidance on the amount of human involvement required, no settled position on joint ownership between a human and an AI developer, and no dedicated framework for the training-data question beyond the general fair-dealing exception that Section 52 was written for entirely different purposes. Calls for a formal legislative update, whether through a standalone AI and copyright framework or a targeted amendment to the 1957 Act, have grown louder as more disputes reach the courts, and the ANI ruling is likely to accelerate that conversation rather than settle it.<\/p>\n

Conclusion<\/u><\/strong><\/p>\n

The short version is that Indian law, as it stands today, does not allow an AI system to own anything, and it does not automatically hand ownership to whoever is using the tool. It asks for a human who can be identified as having caused the work to exist, through prompts, choices, edits, or direction, and the more human judgment is visible in the final product, the stronger the claim to authorship becomes. The RAGHAV episode showed how uneasy Indian institutions are with the idea of a machine as co-author, and the ANI versus OpenAI dispute is now testing the other side of the same problem, whether the data used to build these tools was ever fair game to begin with. Neither question has a final, settled answer yet, and given how quickly generative AI is moving, it may be several years and several more court rulings before one arrives. Until then, anyone creating or building a business around AI-generated content in India would do well to document their own creative contribution carefully, because that is likely to remain the strongest ground to stand on, no matter which way the law eventually settles.<\/p>\n

References<\/u><\/strong><\/p>\n

    \n
  1. Copyright Act, 1957, Section 2(d), Government of India.<\/li>\n
  2. Copyright Act, 1957, Section 52(1)(a), Government of India (fair dealing exception).<\/li>\n
  3. Sahni, Ankit \u2014 RAGHAV Artificial Intelligence Painting App, “Suryast” registration matter, Indian Copyright Office.<\/li>\n
  4. Thaler v. Comptroller-General of Patents, Designs, and Trade Marks, England and Wales High Court (DABUS case).<\/li>\n
  5. Copyright, Designs and Patents Act, 1988, Section 9(3), United Kingdom.<\/li>\n
  6. United States Copyright Office, Compendium of U.S. Copyright Office Practices, guidance on human authorship requirement.<\/li>\n
  7. ANI Media Pvt. Ltd. v. Open AI OpCo LLC, CS(COMM) 1028 of 2024, Delhi High Court, interim order dated July 24, 2026 (Justice Amit Bansal).<\/li>\n
  8. S.S. Rana & Co., “Copy, Right? ANI Media Pvt. Ltd. v. Open AI OpCo LLC,” legal commentary, 2026.<\/li>\n
  9. Chambers and Partners, “Delhi High Court’s ANI v. OpenAI: Shaping AI Copyright Law in India,” 2026.<\/li>\n<\/ol>\n

    \u00a0<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t","protected":false},"excerpt":{"rendered":"

    Author: Nikhil Kumar, \u00a0a 1st year BA.LLB(Hons.) at University of Lucknow Introduction Type a prompt into an AI tool today, and within seconds, you will have a poem, logo, piece of code, or full-length article sitting in front of you. Millions of people in India now use these tools every day for work and fun […]<\/p>\n","protected":false},"author":1,"featured_media":5037,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_themeisle_gutenberg_block_has_review":false,"footnotes":""},"categories":[85],"tags":[],"_links":{"self":[{"href":"https:\/\/lawjurist.com\/index.php\/wp-json\/wp\/v2\/posts\/20643"}],"collection":[{"href":"https:\/\/lawjurist.com\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/lawjurist.com\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/lawjurist.com\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/lawjurist.com\/index.php\/wp-json\/wp\/v2\/comments?post=20643"}],"version-history":[{"count":4,"href":"https:\/\/lawjurist.com\/index.php\/wp-json\/wp\/v2\/posts\/20643\/revisions"}],"predecessor-version":[{"id":20648,"href":"https:\/\/lawjurist.com\/index.php\/wp-json\/wp\/v2\/posts\/20643\/revisions\/20648"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/lawjurist.com\/index.php\/wp-json\/wp\/v2\/media\/5037"}],"wp:attachment":[{"href":"https:\/\/lawjurist.com\/index.php\/wp-json\/wp\/v2\/media?parent=20643"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/lawjurist.com\/index.php\/wp-json\/wp\/v2\/categories?post=20643"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/lawjurist.com\/index.php\/wp-json\/wp\/v2\/tags?post=20643"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}