The Delhi High Court is considering ANI’s copyright case against OpenAI, focusing on LLMs’ access to materials, fair use considerations, and implications for misinformation in AI development.
New Delhi: Restricting Large Language Models (LLMs) like ChatGPT from accessing publicly available copyrighted content may worsen misinformation issues, an Amicus Curiae in ANI’s copyright case against OpenAI told the Delhi High Court on Friday.
Dr. Arul George Scaria, a Professor at NLSIU Bengaluru, argued that LLMs require access to a broad range of materials to function effectively. He stated:
“If I look at the technical literature in this area, the only way to address this information is by giving more access to more materials… in the absence of giving access, they will be trained only on limited sets of materials, which might further misinformation issues.”
Also Read: Umar Khalid’s UAPA Case| Delhi Court Reserves Order on Bail Plea
Dr. Scaria also questioned whether LLMs like ChatGPT could even be developed without access to copyrighted material. Addressing Justice Amit Bansal, he emphasized that AI would play a crucial role in India’s innovation ecosystem.
The Delhi High Court is examining four major legal issues in the case:
- Does OpenAI’s storage of ANI’s news content for training ChatGPT infringe ANI’s copyright?
- Does using ANI’s copyrighted content to generate responses for users constitute copyright infringement?
- Can OpenAI claim ‘fair use’ protection under Section 52 of the Copyright Act, 1957?
- Does the Indian judiciary have jurisdiction over the case, given OpenAI’s servers are based in the U.S.?
During the hearing, the Court noted:
“There is nothing, no sort of law, absolutely in respect of these kind of generative AIs.”
Also Read: Umar Khalid Seeks Bail in North East Delhi Violence Case
ANI alleges that OpenAI is exploiting its content for commercial gain and training ChatGPT without permission. The Court had earlier appointed Dr. Scaria and Advocate Adarsh Ramanujan as amici curiae.
Dr. Scaria submitted that temporary or permanent storage of copyrighted content by LLMs is permissible under law if it is for learning purposes. He highlighted that machine learning should not be treated differently from human learning.
On the issue of OpenAI’s use of ANI’s content, Scaria argued that similar outputs from ChatGPT should not automatically be deemed infringement and that commercial use does not necessarily disqualify an entity from fair use protections.
Also Read: Umar Khalid Bail Hearing: Counsel Argues No Terror Case Established
In contrast, Amicus Adarsh Ramanujan took a stricter stance, stating:
“An act of infringement stands completed when they copy, unless they show that it comes under fair use.”
Ramanujan further stated that the Court should not take an overly policy-driven approach.
The Court will continue hearing arguments on March 10, with further submissions expected from Amicus Ramanujan and ANI’s counsel.
Case Title – ANI Media Pvt Ltd v. Open AI Inc & Anr
FOLLOW US FOR MORE LEGAL UPDATES ON YOUTUBE