IndiaMart Inter Mesh Limited v. Open AI Inc. & Ors. | IP-COM/57/2025 | Calcutta High Court | Justice Ravi Krishan Kapur | May 20, 2026
Kolkata, May 2026 — In a landmark ruling at the intersection of artificial intelligence, intellectual property, and digital commerce, the Calcutta High Court has dismissed an application for interim relief filed by IndiaMart Inter Mesh Limited — one of India’s largest B2B e-commerce platforms — against Open AI Inc., the company behind the globally popular generative AI tool ChatGPT. Justice Ravi Krishan Kapur held that IndiaMart had failed to demonstrate any legally recognisable right that had been infringed, and that no private enterprise can be compelled to promote or display another party’s business on its platform merely because it is commercially convenient to that party.
The Dispute — Selective Exclusion or Legitimate Business Decision?
IndiaMart’s grievance was specific and pointed: when users queried ChatGPT for products available on IndiaMart, the AI tool would bypass IndiaMart’s platform links and instead provide direct links to individual sellers’ websites — while simultaneously providing full platform-level links for competing B2B platforms such as DHGate, Pinduoduo, Shopee, and Taobao. IndiaMart alleged that this constituted deliberate and intentional exclusion of its platform, causing significant loss of user traffic and business revenue. The petitioner characterised this as disparagement, dilution of its registered trademarks, unfair trade practice, and a violation of its fundamental rights under Articles 14, 19, and 21 of the Constitution of India.
The reason for this exclusion, as it emerged during the proceedings, was that IndiaMart’s name appears on the United States Trade Representative’s (USTR) Review of Notorious Markets List 2024 — a list that Open AI had relied upon as an internal policy basis for restricting IndiaMart’s visibility on ChatGPT. IndiaMart contested this, arguing that the USTR List is a foreign document with no statutory force in India, and that blind reliance on it to discriminate against an Indian platform was impermissible — particularly when other platforms on the same USTR List continued to receive full visibility on ChatGPT.
The Court’s Ruling — No Legal Right, No Cause of Action
Justice Kapur dismissed the application on multiple grounds, each of which carries significant implications for the evolving law around AI platforms in India.
On the most fundamental question, the Court held that IndiaMart had no legally enforceable “right to visibility” on ChatGPT — whether arising from contract, statute, or constitutional law. The loss complained of was pure economic loss — a potential reduction in user traffic translating into a potential loss of profit. Such loss, the Court held, does not by itself give rise to a cause of action. Quoting the Delhi High Court’s ruling in Google LLC v. DRS Logistics (P) Ltd. (2023), Justice Kapur observed: “No third party can compel a service provider to use its service in a manner to reflect its link or for its benefit.” Open AI’s decision to rely on the USTR List was characterised as an internal policy and business decision — one that courts cannot be called upon to second-guess in the absence of a demonstrated legal wrong.
On the intellectual property claims, the Court was equally unsparing. It found no case of trademark disparagement, as disparagement requires publication of a false statement — and silence or omission cannot constitute disparagement. On trademark dilution under Section 29(4) of the Trade Marks Act, 1999, the Court held that mere referential use of the IndiaMart name does not satisfy the requirement of “use in the course of trade.” On copyright, the Court noted that no specific infringement of any copyrighted work had been pleaded or identified with particulars — arguments made from the Bar were entirely dehors the pleadings.
The AI Classification Question — Originator or Intermediary?
Perhaps the most intellectually significant portion of the judgment addresses a question that no Indian court has squarely confronted before: is ChatGPT an “intermediary” or an “originator” under the Information Technology Act, 2000?
The distinction matters enormously. An intermediary enjoys conditional immunity — “safe harbour” — under Section 79 of the IT Act from liability for third-party content, provided it observes due diligence. An originator, by contrast, is the source of the electronic message and bears primary responsibility for it. IndiaMart had argued that ChatGPT is an intermediary and is therefore bound by the non-discrimination obligation under Rule 3(1)(n) of the IT Rules, 2021. Open AI countered that ChatGPT is an originator — it generates new content rather than merely transmitting existing content.
Justice Kapur acknowledged that this is a “complicated and vexed question of both law and fact” that can only be finally determined after technical and expert evidence is led at trial. However, for the purposes of the interim application, the Court offered a considered prima facie view: ChatGPT, by virtue of its generative qualities — its ability to synthesise, curate, and produce entirely new content in response to user prompts, going well beyond the function of a search engine or a passive conduit — prima facie falls within the definition of an “originator” rather than an “intermediary.” The Court noted that unlike a search engine, which merely crawls and ranks existing web content, a Large Language Model applies independent algorithms to vast datasets and generates direct, synthesised responses. It can write poems, prepare research projects, generate images — functions that require going beyond existing electronic records and creating new content.
The Court was careful to flag the deeper legislative gap: the IT Act was drafted in 2000, long before generative AI existed. Its definitions of “intermediary” and “originator” were conceived in a world where only humans or legal entities could originate messages. Generative AI does not fit neatly into either category, and the Court expressly called for legislative intervention — not merely government advisories — to create a distinct liability framework for AI platforms that recognises and distributes liability across developers and users in proportion to their control over the system.
Balance of Convenience — Against IndiaMart
Even on the question of balance of convenience, the Court found against IndiaMart. Any interim order directing ChatGPT to display IndiaMart’s links would, in effect, be compelling specific performance of a positive obligation — requiring continuous judicial supervision and being virtually impossible to monitor. The Court also noted a significant factual circumstance: at an earlier point in time, IndiaMart itself had blocked ChatGPT from crawling its website — a fact that considerably undermined its claim of irreparable injury from reduced visibility.
The Ripple Effect — Does This Judgment Unsettle GST on AI Services Too?
The judgment’s implications do not stop at the doors of the IP court. A pointed observation buried within it — that ChatGPT is not a search engine but a generator of entirely new content — quietly opens a significant question in Indian tax law: does ChatGPT, and generative AI more broadly, qualify as an Online Information Database Access and Retrieval (OIDAR) service under the GST framework?
The question matters because under GST, foreign OIDAR service providers are required to register in India and pay GST on their B2C transactions — i.e., services supplied directly to individual consumers who are not registered businesses. Platforms like ChatGPT and Claude have, by default, been treated as OIDAR services and have been expected to comply with this obligation. But the Calcutta High Court’s characterisation of ChatGPT as a creator rather than a retriever puts that assumption under strain.
The word “OIDAR” is not merely a label — it carries a precise definitional meaning. It refers to services involving access to or retrieval of information or a database. Both “access” and “retrieval” presuppose something pre-existing — a stored record, a queryable database, an archived piece of content waiting to be fetched. That is precisely what a search engine like Google does: it crawls, indexes, and retrieves pre-existing web content in response to a query. But that is not what a Large Language Model does. A generative AI does not retrieve your answer from a database — it synthesises it, token by token, in real time, producing a response that did not exist before you asked the question.
The Finance Act 2023 did broaden the OIDAR definition by removing the earlier requirement of “minimal human intervention” — an amendment that was widely understood to bring more digital services within the GST net. But crucially, it left the words “access” and “retrieval” untouched. The definitional gap, therefore, remains very much alive.
If a well-advised taxpayer were to press this argument in a tax dispute, the consequences could be considerable. For B2B transactions, the impact is limited — registered Indian businesses are already required to pay GST on reverse charge, regardless of the OIDAR classification. But for B2C transactions, the stakes are higher: if generative AI services fall outside the OIDAR definition, the obligation on foreign AI providers to register in India and collect GST from individual users may simply not apply. That would represent a significant revenue collection gap — and, more troublingly, a structural arbitrage against Indian digital service providers who are fully subject to GST on equivalent services.
The deeper issue is one of conceptual architecture. OIDAR was designed for a retrieval economy — search engines, streaming platforms, database downloads. Generative AI is a creation economy. The law was built for a world where digital services delivered pre-existing content; it has not yet reckoned with a world where the service itself is the act of creation. The Calcutta High Court has, perhaps inadvertently, handed tax practitioners a powerful new argument. Interesting times lie ahead.
Why This Judgment Matters
This is the first Indian High Court ruling to substantively engage with the legal status of a generative AI platform under the IT Act framework, and its implications extend well beyond the parties before the Court. Three takeaways stand out for businesses, technologists, and policymakers alike.
First, there is no “right to algorithmic visibility.” Businesses cannot claim a legal entitlement to appear in the results generated by an AI platform — or any private digital platform — merely because their exclusion causes them economic harm. The law of pure economic loss does not recognise such a claim in the absence of a specific legal duty owed by the defendant to the claimant.
Second, the USTR List, while not a legally binding document in India, can inform the internal policies of private platforms operating in India. Courts will not ordinarily interfere with such internal policy decisions in the absence of a demonstrated violation of a substantive legal right.
Third, and most significantly, India urgently needs a legislative framework for AI liability — and, as this judgment now suggests, AI taxation. The IT Act’s binary of “intermediary” and “originator” is ill-equipped to handle generative AI, and the GST framework’s OIDAR definition may be equally unprepared. Until Parliament acts on both fronts, courts and tribunals will be left to navigate this terrain case by case — a process that is neither efficient nor conducive to legal certainty for an industry growing at extraordinary speed.
The suit itself has been directed to proceed to an expeditious final hearing, where these questions — particularly the intermediary/originator classification — will be decided conclusively on the basis of technical and expert evidence.
With courtesy to Mr. Manish Sachdeva.
Leave a Reply