Delhi High Court refuses stay on copyright ruling in ANI-OpenAI case
The Delhi High Court on Tuesday refused interim relief to news agency ANI, which sought a stay on a July single-judge ruling that using copyrighted works to train large language models does not amount to copyright infringement. A division bench of Justices Avneesh Jhingan and Manmeet Pritam Singh Arora orally said it would not pass anything ex-parte without the other side. It sought OpenAI's response and will hear the matter on December 8. ANI sued OpenAI in 2024.
Source
Indian Express — Cities · read the original report ↗
Desk check · compared with the source
What the desk checked (5)
- Delhi High Court refused interim relief to ANI on Tuesday and sought OpenAI's response, with the next hearing on December 8. — Attributed in source to a division bench of Justices Avneesh Jhingan and Manmeet Pritam Singh Arora; dates appear in source.
- A single judge held in July that using copyrighted works to train LLMs is fair use and amounts to private use, including research. — Described in source as the single judge's verdict, with quoted wording from the ruling.
- ANI had offered OpenAI a licence to its content for USD 7.5 million before moving court. — Figure appears in source; no document cited, stated as background.
- OpenAI told the court in November 2024 that it had blocklisted ANI's domain from future training. — Attributed in source to OpenAI's submission before the court.
- ANI's counsel Sidhant Kumar Marwah argued fair dealing is a closed, purpose-specific exception under the Copyright Act. — Attributed to named counsel in source.
Analysts’ view opinion
ఇది కోర్టు గదిలో జరుగుతున్న వ్యాజ్యం అయినా, దాని రాజకీయ బరువు తక్కువ కాదు — కృత్రిమ మేధ సంస్థలకు భారతీయ కంటెంట్‌పై లైసెన్స్ ఒప్పందాలు అవసరమా కాదా అనే ప్రశ్న ఇప్పుడు న్యాయస్థానం చేతిలో ఉంది, కానీ దాని పరిష్కారం చివరకు విధాన నిర్ణయ వేదికకే చేరే అవకాశం ఉంది. మంగళవారం ధర్మాసనం స్టే నిరాకరించడం విషయ పరమైన తీర్పు కాదు; ఎదుటి పక్షం వాదన వినకుండా ఏకపక్ష ఉత్తర్వు ఇవ్వబోమనే ప్రక్రియాపరమైన జాగ్రత్త మాత్రమే. అయినప్పటికీ, ప్రస్తుత స్థితి ఓపెన్ఏఐకి కాలయాపన రూపంలో ఒక వ్యూహాత్మక ప్రయోజనం ఇస్తోంది, వార్తా–కంటెంట్ పరిశ్రమకు ఒత్తిడి పెంచుతోంది.
- సింగిల్ జడ్జి తీర్పు అమలులో కొనసాగడం అంటే, డిసెంబర్ 8 వరకు లైసెన్సింగ్ ఒప్పందాలు లేకుండానే ఎల్ఎల్ఎం శిక్షణ కొనసాగే స్థితి కొనసాగుతుంది — ఇది చర్చల బలాన్ని ఏఐ సంస్థల వైపు మొగ్గించే అంశం.
- ఏఎన్ఐ వాదన కేవలం ఒక సంస్థ ప్రయోజనం కాదు; భారత కాపీరైట్ చట్టంలోని 'ఫెయిర్ డీలింగ్' పరిమిత మినహాయింపును అమెరికా 'ఫెయిర్ యూజ్' తరహాలో విస్తరించకూడదనే వాదన, దేశీయ మీడియా–ప్రచురణ రంగం మొత్తానికి ఉమ్మడి ఆసక్తి ఉన్న అంశం.
- ఈ కేసు 'భారతీయ సృజనకారుల హక్కులు' వర్సెస్ 'ఏఐ ఆవిష్కరణలకు అనుకూల వాతావరణం' అనే రెండు రాజకీయంగా ఆకర్షణీయమైన నినాదాల మధ్య ఘర్షణగా మారే అవకాశం ఉంది; ఏ ప్రభుత్వానికైనా ఇది సున్నితమైన సమతుల్య సమస్య.
- కోర్టు తీర్పులు కంటెంట్ యజమానులకు వ్యతిరేకంగా స్థిరపడితే, పరిహారం చట్టపరమైన సవరణ లేదా లైసెన్సింగ్ ఫ్రేమ్‌వర్క్ రూపంలో శాసన వేదికకు మారే డిమాండ్ బలపడవచ్చు.
- ఓపెన్ఏఐ వైపు నుంచి 'ఆప్ట్-అవుట్ సదుపాయం, ఏఎన్ఐ డొమైన్ బ్లాక్‌లిస్ట్' వంటి స్వచ్ఛంద చర్యలను ప్రస్తావించడం — నియంత్రణను నివారించి స్వీయ-నియంత్రణను ముందుకు తెచ్చే సుపరిచిత వ్యూహ శైలిని సూచిస్తుంది.
What to watch — డిసెంబర్ 8 విచారణలో ఓపెన్ఏఐ సమాధానం తర్వాత ధర్మాసనం మధ్యంతర రక్షణపై ఏ వైఖరి తీసుకుంటుందో, అలాగే ఇతర వార్తా–ప్రచురణ సంస్థలు ఈ అప్పీలుకు మద్దతుగా కలిసి రావడమో లేదా సమాంతర వ్యాజ్యాలు దాఖలు చేయడమో జరుగుతుందా అనేది కీలకం.
స్టే నిరాకరణ అంటే సింగిల్ జడ్జి తీర్పును ధర్మాసనం సమర్థించిందని కాదు, మూల వ్యాజ్యం ఇంకా తేలలేదు; అలాగే ఈ కేసుపై ప్రభుత్వం, రాజకీయ పార్టీలు లేదా ఇతర మీడియా సంస్థల వైఖరి ఏమిటనేది ఈ కథనం ఏమాత్రం స్థాపించలేదు.
Deep dive
Research brief · 8 facts · 5 dates · exam-readyThe brief
Context
News agency ANI sued OpenAI Inc and OpenAI OpCo LLC in 2024 in the Delhi High Court — the first case of its kind in India — seeking an injunction against the use of its copyrighted news content to train large language models (LLMs). In July, a single judge refused an interim injunction and took a broad view of "fair use", holding that training LLMs on ANI's literary works amounted to "private" use including research, and would not substitute ANI's market. ANI appealed that verdict, and on Tuesday a division bench declined to stay the single-judge ruling, seeking OpenAI's response instead. The larger suit itself is still to be adjudicated.
Key facts
- The Delhi High Court on Tuesday refused interim relief to ANI, which had sought a stay on the July single-judge ruling that training LLMs on copyrighted works is not copyright infringement.
- The division bench of Justices Avneesh Jhingan and Manmeet Pritam Singh Arora orally said the court "would not like to pass anything ex-parte without the other side".
- The bench sought OpenAI's response and will hear the matter next on December 8.
- ANI filed its suit in 2024 against OpenAI Inc and OpenAI OpCo LLC — the first of its kind in Indian jurisdiction — seeking an injunction against use of its content to train LLMs.
- The single judge held that using ANI's original literary works to train LLMs would amount to "private" use, including research, and qualify as fair use.
- The single judge ruled such use "would not result in market substitution of ANI's works and, consequently, will not affect its market share or cause actual or potential damage."
- Before moving court, ANI had offered OpenAI a licence to its digital video, imagery, photographic and news content for a fee of USD 7.5 million.
- OpenAI told the court in November 2024 that it had already "blocklisted" ANI's domain, excluding it from future training of OpenAI's software.
Timeline
- 2024ANI files suit in Delhi High Court against OpenAI Inc and OpenAI OpCo LLC seeking an injunction against use of its copyrighted content to train LLMs.
- November 2024OpenAI tells the court it has already "blocklisted" ANI's domain, excluding it from future training of its software.
- July (2025)Single judge of the Delhi High Court refuses interim injunction, holding LLM training on ANI's works is fair use and causes no market substitution.
- Tuesday (recent)Division bench takes up ANI's appeal briefly, refuses interim relief, seeks OpenAI's response.
- December 8Next hearing in the appeal before the division bench.
Who has a stake
- ANI (news agency) — Claims statutory copyright over its news content; seeks injunction, licensing revenue (it had sought USD 7.5 million) and remedies for verbatim replication and false attribution by ChatGPT.
- OpenAI Inc and OpenAI OpCo LLC — Ability to train LLMs on Indian-origin content without licensing agreements; says models lose access to original training data after a cut-off date and that ANI's domain is blocklisted.
- Delhi High Court (single judge and division bench of Justices Avneesh Jhingan and Manmeet Pritam Singh Arora) — Deciding the first Indian case on whether LLM training on copyrighted works is infringement or fair dealing.
- Indian news publishers and content creators — The July ruling, if upheld, lets AI developers train on Indian-origin content without licensing deals, affecting their bargaining power and revenues.
- Sidhant Kumar Marwah (ANI's lawyer) — Argues fair dealing is a closed, purpose-specific exception and that public interest in AI growth cannot displace statutory rights.
Why it matters
This is India's first litigation testing whether feeding copyrighted journalism into AI training data is infringement, and the outcome will shape whether Indian publishers can charge AI firms for their archives. The single judge's expansive reading of "fair use" — which ANI says imports the American doctrine into a statute that does not contain it — could set the baseline for every future AI-content dispute in India. With no stay granted, that reading stands while the appeal and the underlying suit remain pending.
UPSC angle
Prelims pointers
- ANI v. OpenAI (2024, Delhi High Court) is the first case in Indian jurisdiction on use of copyrighted content to train LLMs.
- The July single-judge ruling held LLM training on ANI's works to be "private" use including research, and therefore fair use.
- Division bench hearing the appeal: Justices Avneesh Jhingan and Manmeet Pritam Singh Arora; next hearing December 8.
- ANI had offered OpenAI a content licence for USD 7.5 million before moving court.
- OpenAI told the court in November 2024 that ANI's domain had been "blocklisted" from future training.
- ANI's appeal argues "fair dealing" under the Copyright Act is a closed, purpose-specific exception, unlike the wider US "fair use" doctrine.
Mains framing
The ANI–OpenAI dispute crystallises a gap between an analogue-era copyright statute and generative AI's data-hungry training pipelines. The Delhi High Court's single judge read the statutory exception broadly — treating LLM training on ANI's literary works as private use including research, and finding no market substitution or actual or potential damage — a reasoning ANI challenges as importing the wider American fair use doctrine into an Indian statute where fair dealing is a closed, purpose-specific exception. ANI's further grievances go beyond ingestion: verbatim replication of its copyrighted material in ChatGPT outputs and hallucinatory responses falsely attributed to it, which raise attribution and reputational questions distinct from training. OpenAI's position that its models cannot access original training data after a cut-off, and that ANI's domain was blocklisted in November 2024, illustrates how technical remedies and opt-out mechanisms are being offered in place of licensing — the very asymmetry ANI's USD 7.5 million licence offer sought to correct. The way forward, as the litigation proceeds to December 8 and the suit awaits adjudication, lies in appellate clarity on the scope of fair dealing and, legislatively, in settling whether training data use requires consent, remuneration or a statutory licence, while balancing AI innovation against creators' statutory rights.
Key terms
- Large Language Model (LLM)
- An AI system trained on very large volumes of text, such as those behind OpenAI's ChatGPT.
- Fair dealing / fair use
- Statutory exception permitting limited use of copyrighted work; ANI argues India's fair dealing is a closed, purpose-specific list unlike the broader US fair use doctrine.
- Market substitution
- Test of whether the alleged infringing use displaces demand for the original work; the single judge found no such substitution of ANI's works.
- Ex-parte order
- An order passed hearing only one side; the division bench said it would not pass anything ex-parte without the other side.
- Blocklisting a domain
- Excluding a website's content from future model training; OpenAI said it had blocklisted ANI's domain in November 2024.
- Hallucination (AI)
- Fabricated model output; ANI complained of ChatGPT responses falsely attributed to it.
Practice questions
- Does training a generative AI model on copyrighted news content amount to infringement under Indian copyright law? Discuss with reference to the ANI–OpenAI litigation in the Delhi High Court.
- "India's fair dealing exception is closed and purpose-specific, unlike the American fair use doctrine." Examine this claim and its implications for AI development in India.
- Beyond the legality of training data, what remedies should the law provide for verbatim reproduction and false attribution in AI outputs? Discuss.
Grounded only in the source report — figures and dates are the source's, not inferred.