Delhi High Court denies ANI injunction against OpenAI | Magica
Delhi High Court lets OpenAI continue as ANI copyright case moves on
Editorial Team
••📖7 min read
The Delhi High Court has denied ANI Media's request to immediately stop OpenAI from using its news material to train the models behind ChatGPT. Its prima facie fair-dealing finding gives OpenAI interim relief, but leaves the copyright suit, the evidence and the question of compensation unresolved.
The Delhi High Court denied ANI Media's request to immediately stop OpenAI from using ANI material in the models behind ChatGPT.
Justice Amit Bansal found, at the interim stage, that the training use was prima facie fair dealing and that ANI had not shown substantial copying in the challenged outputs.
The ruling keeps the service running, but it does not finally decide whether OpenAI infringed copyright or whether ANI is entitled to damages.
The Delhi High Court has given OpenAI interim relief in India's first substantive judicial examination of AI training and copyrighted news content. Justice Amit Bansal, a Delhi High Court judge, dismissed ANI Media Pvt. Ltd.'s application for an interim injunction against OpenAI Inc. and OpenAI OpCo LLC, finding prima facie that storing ANI's publicly available works to train the large language models underlying ChatGPT fell within the Copyright Act's fair-dealing exception. The court's ruling from the hearing also found no prima facie substantial similarity between the outputs ANI relied on and the agency's reporting.
OpenAI is the U.S.-based AI company that developed ChatGPT, the generative-AI chatbot at the centre of the dispute. ANI is a news agency whose case concerns its original reports: it alleged that OpenAI copied and stored them for model training, that ChatGPT reproduced them for users, and that the chatbot sometimes falsely attributed invented reports to ANI. The suit began in November 2024.
The decision is consequential because an injunction would have required action before the evidence was fully tested. It is not the final answer to whether particular use of ANI's works infringes copyright. The main suit continues, and the court's findings are expressly prima facie—its preliminary view for deciding interim relief.
The Delhi High Court’s conclusion dismissing ANI’s application for an interim injunction, July 24, 2026. Source: LiveLaw.
Two alleged forms of copying, neither proved enough for an injunction
ANI put two different theories before the court. The first concerned the input side: the electronic storage of its articles during training. The second concerned output: whether ChatGPT responses reproduced protected expression from its reports.
Sundar Pichai said Google supports an industry letter on open models, aligning the company with developers and technology firms resisting broad limits on Chinese AI systems while leaving allegations of IP theft and national-security risks unresolved.
Editorial Team
On the first, the court accepted that electronic storage of a literary work can amount to reproduction under the Copyright Act. It nevertheless held, on the current record, that the storage was fair dealing under Section 52(1)(a), which covers private or personal use including research. Its reasoning was specific: the material was used inside a closed training system, was not available to people for access or download, and was analysed and converted into machine-readable inputs to improve the model. The court interpreted “research” in light of technological change and did not treat OpenAI's commercial status alone as disqualifying. The account of the order says the court described that process as research undertaken at humans' behest and for their benefit.
On the output claim, the court drew the conventional line between a news report's protected expression and the underlying events it describes. The cited responses conveyed similar facts in different language and sometimes added commentary, the court found. ANI had used detailed prompts seeking the “exact” wording of material; even then, the court said, the examples did not show substantial reproduction.
The particular examples also narrowed what those tests could establish. ANI's cited articles were published in August and September 2024, while the training cut-offs reported to the court were April 2022 for GPT-4 and April 2024 for GPT-4o. The court therefore said those articles could not demonstrate memorisation by those models. The report on the decision also notes that the court did not treat free online availability as surrendering ANI's copyright.
That is a meaningful constraint on the ruling. It does not say that a different dataset, a response that substantially copied an article, or evidence of a different market effect would necessarily receive the same result.
Why the court favoured damages over a temporary ban
For interim relief, the court also weighed harm from granting or refusing an injunction. It found no material at this stage showing that OpenAI's conduct had cost ANI subscribers, advertising income or news-syndication revenue. ANI's October 2024 proposal to license its digital media, video, photographs and news content to OpenAI for $7.5 million was relevant here—not as a judicial valuation of the works, but as evidence that the agency's claimed loss could be quantified and compensated with money if it ultimately won.
The balance ran the other way for an emergency restraint, the court said. It reasoned that large-language-model development depends on public-domain information and that requiring licences from multiple sources would make it economically unviable. The order also pointed to the potential effect on Indian LLM development and on millions of ChatGPT users in India, many of whom are not paid subscribers. The court's reasoning as reported was about the comparative harm of an interim remedy, not a calculation of the market-wide cost of licences.
The court rejected OpenAI's jurisdiction objection as well. OpenAI had argued that its training and servers were outside India. But ANI's registered and principal office is in Delhi, ChatGPT is offered to Indian users, and the alleged chain of harm was connected to India, the court held. Reporting on the judgment says the case has been returned to the roster process to determine the bench that will hear the continuing suit.
A judicial ruling alongside an unresolved policy choice
The case drew in more than its original parties. The Federation of Indian Publishers, the Digital News Publishers Association and the Indian Music Industry intervened in support of ANI; Flux AI Labs, the Indian Governance and Policy Project and the Broadband India Forum broadly supported OpenAI. That alignment shows that the case concerns a wider dispute over the terms on which protected material can be used as an input to AI systems.
The interim order also sits beside a different policy proposal. A committee working paper associated with the Department for Promotion of Industry and Internal Trade proposed a mandatory blanket licence: developers could train on lawfully accessed copyrighted works but would pay statutory royalties to rightsholders. The working paper did not decide whether existing fair dealing law already covered training. The court's case-specific interim conclusion therefore does not eliminate the possibility that legislation or a final judgment could set a different licensing rule.
Nor did the court conclude that publishers have no tools outside this suit. It noted that ANI could have blocked web crawlers, while OpenAI said it had blocklisted ANI's website from future training crawls and from ChatGPT search and retrieval-augmented-generation functions. Those statements describe the positions and measures considered in this dispute; they do not establish a general technical or legal remedy for every publisher.
What the continuing case must establish
The immediate decision is whether ANI seeks appellate review of the refusal of interim relief; ANI said after reviewing the judgment that it would consider its legal options, including possible appeals. The larger question remains for the case that continues before the Delhi High Court.
For ANI to move beyond the present record, the evidence will need to make the disputed mechanisms and effects more concrete:
whether particular ANI works entered relevant datasets or persisted through development of the models at issue;
whether identified ChatGPT outputs substantially reproduce ANI's original expression, rather than facts or quotations that the agency does not own; and
whether ANI can show compensable market harm, including lost subscriptions, syndication revenue or licensing income.
The ruling gives OpenAI time to operate while that record is developed. Its durable significance will depend less on the court's interim label of AI training as research than on whether later evidence changes the fair-dealing analysis—or whether lawmakers decide that licensing, rather than litigation, should allocate the value of copyrighted inputs.
Brazilian courts and bar associations have responded to hidden prompts in legal filings with fines, referrals, suspensions and an investigation. The known cases were detected, however, and do not show that an AI system changed a judicial decision; they instead test how courts secure AI-assisted review of untrusted case documents.
Salamanca City Central School District has put its proposed Realbotix humanoid robot and Optio AI assistant pilot on hold while it works on student-data privacy agreements. The pause leaves the district with a more basic task: establish what the digital tutor and the physical robot each collect, do and improve before either reaches students.
Samsung Electronics and Broadcom have signed an MOU covering memory, foundry technologies and advanced packaging for next-generation AI infrastructure. The companies expect more than $200 billion of collaboration over five years, but have not disclosed how that figure translates into purchases, production or capacity commitments.
OpenAI says GPT-5.6 Sol reached a 24.9% ExploitGym pass rate under a two-hour cap, up from GPT-5.5’s 15.1%. The benchmark measures a bounded step from a known vulnerability to an exploit; a separate assessment performed with OpenAI finds continued limits on hardened targets and long operations.
Chey Tae-won, chairman of SK Group, said Anthropic had sought SK Hynix component-supply cooperation for chips it hopes to make itself. But Anthropic has disclosed no chip plan, and the same summit also produced a separate letter of intent between Anthropic and SK Telecom on a large AI data center.
South Korea announced $950 billion in planned five-year semiconductor cooperation involving Samsung, SK Group and global technology companies, alongside prospective AI-data-center projects. The figure aggregates arrangements with different terms and disclosed commitments, leaving delivery, power and the overlap between the SK and NVIDIA figures unresolved.
OpenAI said it applied a fix after a July 25 disruption affected ChatGPT, its Codex coding service and parts of its API. The available reporting points to recovery in about 50 minutes, but the retained incident records do not explain the cause, the mitigation or whether the services failed through a common component.
Anthropic CEO Dario Amodei said the Claude developer has signed supply deals with Samsung Electronics and SK hynix. The available disclosures show a useful infrastructure relationship, but not an HBM purchase, a dedicated allocation or a claim on either supplier’s output.
Moody's Ratings expects six US hyperscalers to spend $785 billion in 2026 and close to $1 trillion in 2027. The forecast signals continued demand for AI capacity, but a separate Moody's analysis shows why capex alone understates the financing question: five of the companies already reported $969 billion in undiscounted future lease payments, much of it not yet on their balance sheets.
Anthropic has released Claude Opus 5 at Opus 4.8's API price and says it approaches Claude Fable 5 on coding and knowledge work. An Epoch AI listing places it at 159 ECI, below Fable 5 and GPT-5.6 Sol, leaving customers to test whether the cheaper model delivers better economics in their own workloads.
Nvidia plans to invest $1 billion in Naver as Brookfield signs a nonbinding term sheet for up to $9 billion toward a South Korean AI-factory project. The companies target a 200-megawatt expansion by 2028, but financing, definitive terms and construction remain ahead.
Nvidia and SK Group signed letters of intent for an AI-infrastructure initiative they value at more than $500 billion, centered on an up-to-2-gigawatt Korean AI factory and a long-term SK hynix memory partnership. The disclosed projects give the plan a technical shape, but not a capital breakdown, time horizon or basis for treating the headline figure as committed spending.
Debian has opened discussion of two General Resolution choices: a ban on direct LLM-assisted contributions that would amend its Social Contract, and a conditional-use policy that puts responsibility on contributors. The official text shows that neither policy has been adopted and that the permissive option has changed from its initial mailing-list draft.
Exposed attacker infrastructure contained logs showing the Hermes AI agent running unattended reconnaissance and privilege-escalation checks against systems associated with Thailand's Ministry of Finance. The artifacts show deliberate, operator-directed targeting, but they do not establish the initial access path, the full extent of compromise or data theft.
Amazon is closing its San Francisco AGI Lab amid undisclosed job cuts, while saying its frontier-model research under Pieter Abbeel and its Nova Act service will continue. The closure follows leadership departures and an AGI reorganization; Amazon has not said how many people are affected or where the lab’s work will go.
Delhi Police says its live facial-recognition deployment at Jantar Mantar is meant to find wanted people and prevent disruption at a large student protest. The public record still does not show the current match threshold, the legal basis and scope of the watchlist, how long footage is kept or the safeguards after an alert.
Vice Premier Ding Xuexiang’s visits to Huawei and computing facilities connect China’s domestic AI-chip push to a national effort to schedule compute, data and power. The harder test is whether those systems can make domestic hardware efficient and dependable while Chinese developers and cloud rivals reduce their reliance on Huawei.
A viral reel recorded at a Delhi trans-rights protest with Ray-Ban Meta glasses was initially left up by Instagram and later removed. The episode shows how a recording indicator and user rules can fall short when people filmed without consent must discover the camera, navigate moderation and contend with copies after the fact.
CXMT reportedly charged more than Samsung for comparable DDR5 server modules and signed large supply agreements with ByteDance and Tencent. The reported premium reflects tight supply and state-shaped domestic demand, while CXMT remains behind leading rivals in high-bandwidth memory and lithography capability.
