DataPacts

ANI v. OpenAI (Delhi High Court): Claims, Rulings and Status

Checked

At a glance

ANI v. OpenAI is a copyright suit that the news agency ANI Media Pvt. Ltd. brought against OpenAI, named in the judgment as Open AI OpCo LLC, in the High Court of Delhi in November 2024 (CS(COMM) 1028/2024), claiming that OpenAI stored and trained on ANI's news content and reproduced it in ChatGPT's answers.

Independent reference. Every buyer is covered on the same terms, every figure carries its source and check date, and no buyer pays to be listed or placed higher.

Request an independent introduction

Tell us what data your company holds. We reply with the buyers that fit and what their terms leave out. Your details reach a buyer only after you agree to a named introduction.

IndependentYour details stay with us until you agreeNot selling is a valid outcome

On July 24, 2026 Justice Amit Bansal refused ANI an interim injunction, holding prima facie that storing ANI's works to train the models falls within fair dealing under section 52(1)(a) of the Copyright Act 1957. As of November 9, 2026 the suit is pending.

The ruling decides only the interim application and, in the court's words, has no bearing on the final outcome (para. 274). This page is part of the independent reference on AI training data licensing, DataPacts is not affiliated with either party, and nothing here is legal advice.

All paragraph numbers, dates and docket numbers come from the court's judgment, and how figures are sourced and dated is set out on the methodology page.

ANI v. OpenAI at a Glance: Court, Suit Number, Parties and Status

No court has decided the suit itself: the only reasoned ruling is the interim judgment of July 24, 2026, which refused ANI's injunction on a prima facie view. The table lists the case record as of 2026-11-09 for status, with the judgment re-read on 2026-11-17.

Field Value Source
Case ANI Media Pvt. Ltd. v Open AI OpCo LLC judgment caption
Court High Court of Delhi at New Delhi (Justice Amit Bansal) judgment caption, CORAM
Suit number CS(COMM) 1028/2024; interim application I.A. 45300/2024 judgment caption
Filed November 2024 case record
First listed 2024-11-19 (four issues framed, two amici appointed) judgment para. 28
Plaintiff ANI Media Pvt. Ltd. (Asian News International), New Delhi judgment paras. 6, 45
Defendant Open AI OpCo LLC judgment caption
Claims training claim; output claim (copyright, Copyright Act 1957) judgment para. 6
Interveners 6 (3 broadly for ANI, 3 broadly for OpenAI) judgment paras. 29, 30
Interim ruling 2026-07-24: I.A. 45300/2024 dismissed, prima facie judgment para. 273
Status suit pending, as of 2026-11-09 court record; LiveLaw report of the judgment

What ANI Claims: Training on Its News and Reproduction in ChatGPT

ANI claims copyright infringement on two counts: a training claim, that OpenAI copied and stored ANI's content to train its large language models, and an output claim, that ChatGPT reproduces ANI's works in its answers (judgment para. 6).

ANI is a news agency that licenses its syndicated feed to subscribers on non-exclusive terms that exclude further syndication (para. 244).

ANI argued that public availability of its content does not amount to a license, that tokens and vectors are reproductions, and that OpenAI also infringed its translation right (paras. 57.6, 57.8, 127.2, 127.3).

ANI also pointed to OpenAI's licenses with the Financial Times, the Associated Press and Condé Nast as recognition of news publishers' rights (para. 57.7). The Associated Press announced its license on 2023-07-13, the Financial Times on 2024-04-29 and Condé Nast on 2024-08-20, and none of them disclosed a value.

What ANI Asked the Court to Order Before Trial

ANI's interim application, I.A. 45300/2024 under Order XXXIX Rules 1 and 2 of the Code of Civil Procedure, asked for relief before trial, which OpenAI described as a demand to delete ANI's works "held and stored in whatever form" (para. 258.1).

OpenAI argued that deletion would amount to the final relief and would conflict with its duty under US law to preserve training data (para. 258.1). One amicus read ANI's request that OpenAI obtain a license as showing that the claim could be compensated in money (para. 261.2).

The Four Issues the Court Framed on November 19, 2024

The court framed four issues for the interim application on the first day the suit was listed, November 19, 2024, and appointed two amici curiae the same day (paras. 9, 28). The table gives each issue in shortened form with the court's prima facie answer as of 2026-07-24.

Issue Question as framed (shortened) Prima facie answer Paras
1 Does storing ANI's news data for training ChatGPT infringe? No: storage for training is fair dealing under s.52(1)(a) 9, 256, 271
2 Does using ANI's data to generate responses infringe? No substantial reproduction shown; memorization left to trial 9, 124, 125
3 Does the use qualify as fair use (read as fair dealing) under s.52? Yes: purpose test and fairness test met 9, 219, 255
4 Do Indian courts have jurisdiction although the servers are in the US? Yes: CPC s.20 and Copyright Act s.62(2); no extraterritorial application 9, 48, 55

The court decided jurisdiction (Issue 4) first, then outputs (Issue 2), then storage and fair dealing (Issues 1 and 3) together, because the amici and the parties agreed they were intertwined (para. 10). Issue 3 says "fair use" but the court reads it as "fair dealing", the term of section 52(1)(a) (para. 9).

Six Interveners and Two Amici Curiae

Six organizations intervened, three broadly supporting ANI and three broadly supporting OpenAI, and the court allowed all six on July 4, 2025, limited to the legal issues framed (paras. 29 to 31). The table lists each intervener as of 2026-07-24, with the application number and the judgment's own description.

Intervener Application Description in the judgment Broadly supported Source
Federation of Indian Publishers I.A. 470/2025 national body of the Indian publishing industry ANI paras. 29.1, 30
Digital News Publishers Association (DNPA) I.A. 2199/2025 digital news platforms ANI paras. 29.2, 30
Indian Music Industry (IMI) I.A. 4027/2025 body of phonogram producers ANI paras. 29.3, 30
Flux Labs AI Private Ltd. I.A. 4615/2025 generative AI start-up OpenAI paras. 29.4, 30
IGAP Project LLP I.A. 4616/2025 law and technology think tank OpenAI paras. 29.5, 30
Broadband India Forum I.A. 10766/2025 policy forum and think tank OpenAI paras. 29.6, 30

The amici curiae were Adarsh Ramanujan and Professor Arul George Scaria, appointed on 2024-11-19 (para. 28). They disagreed on jurisdiction over the training claim, as the next section shows.

Can an Indian Court Hear a Claim About Training on US Servers?

Yes, prima facie: the court held that it has territorial jurisdiction under section 20 of the Code of Civil Procedure and section 62(2) of the Copyright Act, and that hearing the training claim does not apply the Indian Copyright Act outside India (paras. 48, 55).

OpenAI argued that training takes place outside India on data stored on servers outside India, that sections 62 and 1(2) of the Copyright Act do not reach abroad, and that ChatGPT's terms of use point to courts in San Francisco (paras. 36.2, 36.3, 36.6).

ANI argued that its registered and principal office is in Delhi, so section 62(2) applies, and that OpenAI targets paying users in Delhi (para. 37).

The amici split: Professor Scaria sided with OpenAI on the territoriality of the training claim, and Mr. Ramanujan treated access and transmission from India as part of it (paras. 41.2, 42.4).

The court called storage on US servers "a terminal step in the chain of events which begin from access of copyrighted works from India" (para. 51). It applied Neetu Singh v.

Telegram to servers abroad (paras. 52, 53) and declined to separate the training and output claims (para. 54). How other countries treat training copies made abroad is covered on the page on copyright and AI training across jurisdictions.

The Interim Judgment of July 24, 2026

The 135-page judgment, reserved on March 27, 2026 and pronounced on July 24, 2026, upheld jurisdiction and dismissed ANI's application I.A. 45300/2024 on outputs, storage and balance of convenience (paras. 271 to 273). The table summarizes the four findings as of 2026-07-24.

Question Court's prima facie finding Key reasoning Paras
Jurisdiction upheld US storage is the terminal step of a chain from India; claims not separated 48 to 55
Outputs no infringement shown examples post-date training cut-offs; retrieval not pleaded; no substantial reproduction 82, 124, 125
Storage for training fair dealing, s.52(1)(a) commercial use not excluded; lawful source; purely private; research by updating construction; fairness factors met 198 to 256
Balance of convenience against ANI unused opt-out; blocked by OpenAI; no lost subscribers; license offer makes the claim quantifiable; public interest 262 to 270

The Output Claim: Articles Published After the Training Cut-Off

The court found no prima facie case on outputs because the articles ANI used as examples were published after the training of the models available to ANI had ended.

The training cut-offs of April 2022 for GPT-4 and April 2024 for GPT-4o were undisputed, and the plaint's nine illustrations of reproduction came from articles published later in 2024 (para. 82). OpenAI counted 14 works published in August or September 2024, which is its own submission (para. 61.1).

The court read the examples as live links "perhaps reflecting RAG technique" and noted that infringement by retrieval-based outputs was not pleaded (para. 84 and following). It found no substantial reproduction and left memorization as a disputed question for trial (paras. 124, 125).

Storage for Training Under Section 52(1)(a): the Purpose and Fairness Tests

The court held prima facie that OpenAI's storage of ANI's works to train its models is fair dealing for "private or personal use, including research" under section 52(1)(a), so it does not infringe under section 51 (paras. 256, 271).

On the purpose test, commercial use does not by itself take a use outside section 52(1)(a)(i) (paras. 176 to 181).

The copies came from ANI's freely available website, without any paywall broken, so they were not infringing copies, and Bartz v. Anthropic was distinguished on shadow libraries (paras. 198, 199, 200).

Storage "in a closed space without access to the public" is "purely private" (para. 211), and "research" covers training by updating construction (paras. 216 to 219).

What this reading means for Indian law in general, the open statutory text and the personal data rules are set out on the page on India's rules for licensing data for AI training.

On the fairness test, the court found the use limited to training (para. 243), no market substitution of ANI's works (para. 251) and a public interest in the technology (paras. 252 to 254).

Balance of Convenience: the Unused Opt-Out and the License Offer

The court also refused the injunction on balance of convenience, irreparable injury and public interest, finding all three against ANI (paras. 270, 272). ANI could block crawlers but had not, and OpenAI stated that it had blocked ANI's website from its crawlers for training and from ChatGPT search and retrieval (paras. 258.4, 262).

How GPTBot and OpenAI's other crawlers work and are blocked is covered on the crawler page. The court found no evidence of lost subscribers (para. 264).

It noted that ANI had offered OpenAI a license for a fee by letter of October 3, 2024, which made the claim quantifiable (para. 265), and that an injunction would harm users and LLM development in India (paras. 266 to 269).

ANI v. OpenAI Hearings, 2024 to 2026

The judgment records 21 dates on which submissions were heard, from February 21, 2025 to March 27, 2026, after the first listing on November 19, 2024 (paras. 28, 33). The table gives the record as of 2026-07-24, from the judgment.

Date Event Source
2024-11-19 first listing; four issues framed; amici appointed para. 28
2025-02-21 to 2025-11-21 17 hearing dates (2025-02-21, 03-10, 03-18, 03-28, 04-02, 04-22, 04-29, 07-04, 07-23, 07-30, 08-05, 08-18, 09-12, 09-23, 10-17, 11-07, 11-21) para. 33
2025-07-04 six interventions allowed para. 31
2026-01-23 to 2026-03-20 hearings on 2026-01-23, 02-20, 03-20 para. 33
2026-03-27 final hearing; judgment reserved; closing submissions permitted para. 33
2026-07-24 judgment pronounced; I.A. 45300/2024 dismissed caption; para. 273

Professor Scaria's closing submissions dated 2026-04-02 were taken on record (para. 34.1). No later order was opened as of 2026-11-17.

Where the Suit Stands After the Interim Judgment

As of November 9, 2026 the suit is pending in the High Court of Delhi, and the interim judgment is the only reasoned ruling on the record. The findings are prima facie and expressly have no bearing on the final outcome (para. 274).

No appeal, appellate bench or appeal hearing is on a court record opened for this page as of 2026-11-17, and no next hearing date or trial schedule is known. The suit is not on the Delhi High Court's advance cause list of 2026-11-17, which was searched, and the court's case-status form could not be used.

What the Case Shows Holders of News Content

The interim judgment treats a holder's own opt-out and its willingness to license as reasons against stopping training before trial.

Under this reading, freely accessible content was copied lawfully for section 52(1)(a) purposes and its storage for training was private. The licensing options of news publishers licensing content to AI companies are set out on the publishers page.

The court weighed ANI's own license offer as a sign that money could compensate ANI (para. 265); the licenses OpenAI holds with other publishers appear only as ANI's argument (para. 57.7).

What news archives as AI training data bring to a training corpus is covered on the news data page.

ANI v. OpenAI is the only Indian case in the tracker and one of several suits against OpenAI, next to the German lyrics judgment and the consolidated US cases. Every tracked case with its docket and status is listed in the case tracker of AI copyright lawsuits over training data.

The table compares three related cases against OpenAI or on training abroad, with status as of 2026-11-09.

Case Court and docket Claim type Status Source
GEMA v. OpenAI Landgericht München I, 42 O 14139/24 copyright (song lyrics, memorization) judgment 2025-11-11 largely for GEMA; not final Munich Regional Court press release 11/2025
In re OpenAI Copyright Infringement Litigation S.D.N.Y., MDL No. 3143 copyright (news, books) cross-motions for summary judgment briefed; no ruling as of 2026-11-09 JPML transfer order
Getty Images v. Stability AI (UK) High Court, IL-2023-000007; Court of Appeal CA-2026-000007 copyright (images); training claim abandoned (no evidence of training in the UK), cited by an amicus here (para. 42.5) High Court judgment 2025-11-04; appeal on secondary infringement permitted judiciary.uk judgment

OpenAI's licenses, crawlers and other disputes are on the profile of OpenAI as a data buyer.

More About ANI v. OpenAI

Decided Versus Still Open in ANI v. OpenAI

Only the interim application is decided, and every finding in it is prima facie.

Decided (interim, prima facie) Still open
jurisdiction upheld memorization at trial
outputs not shown to reproduce ANI's works any final finding on section 52(1)(a)
storage for training is fair dealing damages
I.A. 45300/2024 dismissed on 2026-07-24 any appeal and its outcome

Frequently Asked Questions About ANI v. OpenAI

Did ANI lose its case against OpenAI?

Not the case: ANI lost its interim injunction application on 2026-07-24. The suit is pending as of 2026-11-09.

Only prima facie and only for this application. Storage for training fell within section 52(1)(a), and the court said its findings have no bearing on the final outcome.

Why could a Delhi court hear a case about training done in the United States?

ANI's office is in Delhi (section 62(2)), and OpenAI serves users there (CPC section 20). The court treated US storage as the last step of a chain starting in India.

Has ANI appealed?

No appeal is on a court record opened for this page as of 2026-11-17.

Where can I read the ANI v. OpenAI judgment?

The judgment of 2026-07-24 in CS(COMM) 1028/2024, I.A. 45300/2024, 135 pages, is published as a PDF by LiveLaw.