Stay updated with the latest - Click here to follow us on Instagram
The Delhi High Court on Tuesday sought the response of OpenAI in a suit filed by news agency ANI alleging that the artificial intelligence research organisation used its copyrighted material for AI training purposes. The first-of-its-kind suit in Indian jurisdiction was filed against Open AI Inc (OAI) and Open AI OpCo LLC.
The ANI, in its suit, is tentatively seeking damages of Rs 2 crore and is seeking that the court pass a decree of permanent and mandatory injunction restraining OAI or any person acting on their behalf from storing, publishing, reproducing or in any manner using ANI’s works.
Open AI on the other hand, through its counsel senior advocate Amit Sibal, submitted before the court of Justice Amit Bansal that subsequent to a legal notice issued to the San Francisco-based company by ANI Media Pvt Ltd in September, as of October 2024, OAI has blocklisted ANI’s domain. This, in effect means that it is not using material from the domain for training its large language models (LLMs).
The court also recorded that “given the range of issues involved in the present suit as well as the issues arising on account of latest technological advancements vis-a-vis copyrights of various copyright owners, this court is of the view that an amicus curiae be appointed.” The court is expected to disclose the details of the amicus curiae so appointed later in its publicly available order.
ANI, through its counsel Sidhant Kumar Marwah, has alleged copyright infringement by OAI and has listed “three broad buckets of cause of action”. This includes the use of publicly available copyrighted material of ANI being used by OAI to train their LLMs. A second cause of action, as Marwah argued, is that once a query is put on ChatGPT, it provides “verbatim or substantially similar (content to) my copyrighted material which has been used by you (OAI) to train your software.”
A third aspect is also of hallucinatory responses being provided by ChatGPT, while falsely attributing it to ANI, the agency alleged.
Marwah informed Justice Bansal, “Third aspect, which is most problematic not only from private rights but also public rights perspective is, it gives false attribution…e.g. Rahul Gandhi gave an interview to ANI. In fact no such interview was given…”
Sibal, meanwhile, contested the territorial jurisdiction for adjudication of the matter. Noting that OAI is facing 13 lawsuits in the USA, two in Canada and one in Germany, Sibal submitted that while the software is “a force for good, it is not perfect”.
“No court (globally in the multiple suits OAI is facing) has found prima facie infringement of copyright… I’m always transparent about what I do… There is no monopoly on facts…Copyright does not protect facts, it protects expression… No suit in India, this is the first. I don’t access anything illegally. I have been operating since 2022, I’m not located in India, my servers are located abroad, I do not reproduce any material in India, training also doesn’t happen in India,” Sibal submitted, pointing out that thus ANI has no cause of action or reason to approach a court in the country.
“The way the model works, when a query is made, the software does learn from information from various sources but does not access the database… It is not that ANI’s information is stored on a database,” Sibal argued.
Responding to the aspect of hallucinatory answers, Sibal submitted, “There are certain instances when there are false attributions… These are softwares, they are automated…through manipulation by way of query you can have manipulated answers.”
While Justice Bansal refused to grant any ad interim injunction on Tuesday, the court posted the matter for further consideration on January 28.
Stay updated with the latest - Click here to follow us on Instagram