Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Fix the driver behind crashes, sound loss and screen glitches3Clear out junk files and repair common Windows errorsShort answer: India’s existing Copyright Act was flexible enough for the Delhi High Court to deny ANI’s interim request against OpenAI, but the decision does not make all commercial AI training lawful. In its July 24, 2026 judgment, the court treated OpenAI’s temporary storage of ANI material for large-language-model training as prima facie protected by fair dealing under Section 52(1)(a), while finding the cited ChatGPT outputs insufficiently proven as substantial reproductions. The case exposes major statutory gaps rather than closing them.
The ruling in ANI Media Pvt. Ltd. v. Open AI OpCo LLC is an interim-relief decision, not a comprehensive text-and-data-mining regime. Its reasoning may influence later courts, appeals and legislation, but the outcome remains tied to ANI’s evidence, the pleaded uses and the particular examples before the court.
What the Delhi High Court actually decided
ANI alleged that OpenAI copied and stored its news material to train the models behind ChatGPT without permission, and that ChatGPT later generated responses resembling ANI’s protected expression. The court considered those as separate legal disputes and rejected ANI’s application for an interim injunction. The judgment is available at the reported decision; the Delhi High Court is the authoritative place to check subsequent orders or appeals.
At the interim stage, the court found a prima-facie basis for treating storage of ANI’s literary works for model training as covered by Section 52(1)(a)’s fair-dealing exception. It also found that ANI had not established sufficient memorisation or substantial reproduction in the cited outputs. That is a procedural win for OpenAI, not a final declaration that its entire training corpus complies with Indian law.
#1 Best Overall
One lawsuit contains two different copyright questions
Training and dataset copies
AI development can involve web crawling, downloading, temporary storage, cleaning, tokenisation, dataset preparation and model training. Each step may create or process a copy. The legal question is not answered merely by observing that a finished model does not display every source article as a readable document.
Responses and possible reproduction
A later ChatGPT answer raises a different issue. A response may convey the same facts as a news report without copying its wording. Conversely, a response that reproduces distinctive phrasing, structure, sequence or long quotations could present a stronger infringement case. Training, retrieval from a source database and verbatim output are technically different processes and should not be treated as one event.
| Technical stage | Illustrative legal question |
|---|---|
| Crawling or downloading | Was a protected work copied, and was the acquisition authorised? |
| Temporary storage and processing | Does electronic storage engage the reproduction right, and does an exception apply? |
| Model training and parameters | What information is retained, in what form, and can it be shown to reproduce expression? |
| Retrieval at answer time | Is the system delivering or copying a source document rather than generating a new response? |
| Output publication | Is the answer substantially similar to protected expression, or merely similar in facts or ideas? |
Why servers outside India did not end the case
OpenAI argued that relevant servers were outside the United States. The court nevertheless held that Indian courts could entertain the dispute and then separately examined the substantive copyright issues. A service can be hosted abroad while being offered to Indian users, affecting Indian publishers and markets, or operating through commercial connections with India.
- Jurisdiction asks whether an Indian court can hear the dispute.
- Choice and application of law ask which substantive rules govern the alleged acts.
- Enforcement asks whether an order can practically bind conduct, entities or servers located abroad.
ANI’s success on jurisdiction did not establish infringement. Cross-border AI cases will have to address all three questions.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Rank #2
How Sections 14, 51 and 52 fit together
Section 14: the exclusive rights
Section 14 of the Copyright Act, 1957 gives a copyright owner exclusive rights to perform specified acts, including reproduction of a literary work and electronic storage. The statutory text is available through India Code. In an AI case, ingestion copies, transformed representations and later outputs may involve different applications of that right.
Section 51: the infringement framework
Section 51 supplies the principal infringement framework: identify the exclusive right implicated, ask whether the defendant performed a reserved act, and then test any statutory exception. Ownership and protection of the particular work must also be proved. The ANI judgment did not ignore copying; it considered the reproduction issue and then examined Section 52.
Section 52(1)(a): fair dealing
Section 52 contains exceptions to infringement. The relevant provision refers to fair dealing for specified purposes, including private or personal use and research. The Copyright Office’s exceptions page sets out the statutory framework.
The court’s useful analytical structure separates two questions:
Free tools Windows power users keep installed
One-click scans. No signup required.
- Purpose: does the use fall within a purpose listed in Section 52(1)(a)?
- Fairness: considering all the facts, is the dealing fair?
The court did not automatically disqualify OpenAI because it is a commercial company. Commerciality can still affect fairness, economic harm and market substitution; it is not a blanket exemption or an automatic defeat.
Why the fairness analysis remains fact-intensive
Relevant questions include whether the use was limited to model development, whether sources were publicly accessible, whether the use substitutes for a licensing market, whether protected expression appears in outputs, and whether licensing a very large corpus would be practically workable. Public availability is relevant context, but it does not put a work in the public domain or waive copyright.
The court also considered public interest, including effects on users, innovation and AI development. That interest must be balanced against the cost to publishers whose reporting supplies training material and against the possibility that unlicensed use weakens viable licensing markets.
The DPIIT working paper identified that Indian law has no expressly named generative-AI text-and-data-mining exception. The paper is not legislation. The ANI decision is therefore an interpretation of existing language, not the creation of a purpose-built statutory licence.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Rank #4
Why ANI’s output claim was weaker on the evidence shown
The court reportedly found that the examples relied on by ANI did not establish sufficient memorisation or substantial reproduction. Copyright protects original expression, not facts and ideas in the abstract. A chatbot stating that the same event occurred may be factually accurate without copying a report’s wording or arrangement.
That does not make outputs categorically safe. A different case could involve repeated verbatim passages, a distinctive sequence of paragraphs, prompt-sensitive reproduction or retrieval of a source document. Courts may compare works as a whole rather than isolate a few common fragments. Ownership, assignments and authorship arrangements also matter when a publisher brings the claim.
Is India’s Copyright Act outdated?
It is technologically dated: the Act predates generative AI and does not expressly answer whether training copies, tokenisation, model weights or temporary datasets are infringing. It also does not specify opt-out rights, transparency duties, collective remuneration, proof standards for training-corpus use or remedies for synthetic outputs that compete with a source market.
Yet “old” does not mean unusable. General concepts such as reproduction, fair dealing, substantial similarity, jurisdiction and public-interest balancing allowed the court to decide an interim application. The harder question is predictability at industrial scale. A case-by-case test may produce different outcomes for news, books, music, images, software, databases, pirated material, confidential information and personal data.
What the ruling does not establish
- It does not say that every AI company may scrape anything available online.
- It does not hold that commercial AI training is always fair dealing.
- It does not decide that ChatGPT or another model can never infringe through an output.
- It does not turn public accessibility into a copyright waiver.
- It does not resolve final liability, damages, licensing, deletion, retraining or disclosure remedies.
- It does not automatically extend to paywalled, pirated, leaked, confidential or contract-restricted sources.
Practical implications for publishers and creators
This is general information, not legal advice. A rights-holder assessing a suspected use should:
- Preserve publication dates, source files, authorship records and assignments.
- Capture suspicious outputs with the exact prompt, account, date, URL and screenshots.
- Compare the complete source work, distinguishing copied expression from shared facts.
- Document licensing terms, crawler policies, paywalls and any contractual restrictions.
- Seek Indian copyright advice before demanding a takedown or filing suit, particularly where ownership or cross-border enforcement is contested.
Practical implications for AI companies
Developers should not treat the ANI ruling as a substitute for compliance planning. Sensible controls include:
- maintaining provenance records for licensed, public, user-provided and uncertain sources;
- documenting acquisition and deletion processes;
- testing for memorisation and long, distinctive verbatim outputs;
- separating retrieval systems from model-training pipelines in risk assessments;
- providing complaint, correction and escalation channels;
- preserving logs and dataset evidence relevant to litigation; and
- analysing Indian law separately from United States or European rules.
What would make the framework more predictable?
Courts can resolve individual disputes, but Parliament or regulators are better placed to establish clear licensing and transparency rules. Possible reforms include a defined research or text-and-data-mining exception, rights-holder opt-outs, collective remuneration, disclosure of training sources, safeguards for confidential and personal data, and tailored remedies such as output controls or negotiated licences rather than automatic injunctions.
For businesses, legal-research services such as Westlaw and Lexis may help track decisions where available, while free primary materials remain available at India Code, the Delhi High Court and the Copyright Office. Availability, Indian-law coverage and pricing vary. An enterprise AI subscription, including information at OpenAI Business, does not itself grant permission to ingest a third party’s copyrighted corpus.
PC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchThe bottom line
India’s Copyright Act can address an individual AI dispute: the ANI judgment used existing reproduction and fair-dealing concepts to deny interim relief. But it does not supply the detailed, technically specific and predictable rules needed for modern, large-scale model training. The decisive questions in future cases will be what was copied, how it was stored or retrieved, what the output reproduced, what evidence proves it, and whether the use was fair in its economic and public-interest context.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
