Context: Last week, a copyright infringement lawsuit by India’s Asian News International (ANI) against OpenAI became discoverable in the Delhi High Court (HC) (November 18, 2024 ai fray article). An initial case management conference was held the following day, and OpenAI’s counsel declared that his client had stopped training ChatGPT with ANI’s content as of October (November 19, 2024 ai fray article).
What’s new: The summons issued by the Delhi HC after the hearing has now become publicly accessible. It indicates that ANI is seeking a preliminary injunction (PI), which would be the first in AI-related copyright litigation worldwide. Given the importance and complexity of the dispute, the Delhi HC has appointed two experts (a lawyer and an academic) as amici curiae (“friends of the court”).
Direct impact: The language of the PI sought apparently goes beyond what OpenAI has voluntarily decided not to do. For example, it seeks to bar OpenAI from reproducing any ANI content, which could happen even on the basis of ChatGPT training that occurred prior to October (when OpenAI voluntarily stopped doing so).
Wider ramifications: This case will continue to draw attention to India as a copyright jurisdiction. There will be another hearing in late January.
This is the order (with certain passages highlighted by Saikrishna & Associates‘ Akshat Agrawal, who shared the document and is also the co-author (with Sneha Jain) of a brand new paper on Indian Copyright Law and Generative AI (SSRN link):
Jurisdictional objection raised
Mr. Sibal, whose function in India is comparable to that of a King’s Counsel in the UK, raised a preliminary objection on jurisdictional grounds, given that OpenAI’s servers are located in the United States. The court will have to decide on that one at an early stage.
PI scope sought vs. OpenAI’s cease-and-desist statement
For the record, Mr. Sibal declared:
“[T]he defendants have already bl[a]cklisted the plaintiff’s domain ‘www.aninews.in’ in October, 2024. As a result, the said domain will be excluded from the future training of the defendants’ software.”
Theoretically, a cease-and-desist covenant can moot an injunction request. Not so here. What ANI is seeking goes beyond what OpenAI says it has done to narrow the dispute:
“a. Pass an ex-parte ad-interim injunction restraining the
Defendants, or any person acting for or on their behalf, either
directly or indirectly, from storing, publishing, reproducing or in any
manner using, including through th[e] ChatGPT model, the copyrighted
Plaintiff’s Works or any other original works of the Plaintiff.
b. Pass an ex-parte ad-interim order directing the defendants to
disable access of ChatGPT to Plaintiff’s Works published anywhere
by the Plaintiff or its subscribers.”
Here are some key differences:
- There is apparently some dispute between the parties over whether OpenAI (still) stores ANI’s content or actually just parsed it for training purposes.
- In any event, even if OpenAI can convince the court that it is not “storing” any of ANI’s copyrighted works, there would still be a dispute over whether it is “publishing, reproducing or in any manner using” such material.
- The second part of ANI’s injunction request relates not only to ANI’s own website, but also “its subscribers.” That means any newspaper or other website that pickedup ANI’s news may have served, or may in the future serve, as input for ChatGPT training purposes. The fact that some content is so widely disseminated that an AI system may have obtained it from any number of sources was also discussed at last week’s background briefing by GEMA, a German music rights collecting society suing OpenAI in Munich (November 19, 2024 ai fray article).
Judge, amici curiae and counsel
The judge presiding over this case is Justice Amit Bansal. He has appointed two amici curiae (“friends of the court”), not in the sense of third parties making unsolicited submissions like in the United States, but as independent experts:
- Adarsh Ramanujan, an independent counsel
- Dr. Arul George Scaria, Professor of Law, National Law School of India University (NLSIU)
The legal team representing ANI consists of UNUM Law’s Sidhant Kumar as well as Akshit Mago, Manyaa Chandok (misspelled as “Monyaa” in the order), Om Batra and Anshika Saxena.
OpenAI is represented by Amit Sibal, Khaitan’s Sanjeev Kapoor, Nirupam Lodha, Madhav Khosla, Moha Paranjpe, Gautam Wadhwa, Malika Nandkedlyar, Rajat Bector, Ankit Handa (from Mr. Sibal’s chambers), Darpan Sachdeva and Saksham Dhingra.