OpenAI, which owns the ChatGPT, told the Delhi High Court on Tuesday that the copyright protection in the news reporting is very narrow as there exists a greater public interest in dissemination of information.

The submission was made before the bench of Justice Amit Bansal during the hearing of Asian News International (ANI)'s copyright infringement suit against OpenAI.

"There may be similarity in facts, ideas, names, places but the expression is different. And all that copyright [law] protects is the expression, and much more so in the case of news, the copyright protection is narrow because of all things news is further to be disseminated. There is a greater public interest in dissemination of the ideas and facts that are involved in news reporting. But news reporting has more than ideas and facts, no difficulty. It may be copyrightable, but what is copyrighted is the expression," Senior Advocate Amit Sibal submitted on behalf of ChatGPT.

Sibal added that level of protection in different works will vary depending upon their nature. In this regard, he also gave an example.

"If someone plagiarizes the latest novel of Chetan Bhagat, the test is different because it is fiction, not fact. But where somebody gives their own version of what is narrated in a news article or news report, then the level of protection is much narrower because of twin object - to incentivize creation and promote dissemination of new works," Sibal said.

Sibal was responding to whether the use by OpenAI of ANI’s copyrighted data to generate responses for its users would amount to infringement of the news agency’s copyright.

The question has been framed by the Court to decide ANI's plea seeking interim relief during the pendency of its suit.

In his submissions, Sibal also argued that ChatGPT does not replicate, reproduce or replicate only original works. Sibal further said that for the same prompt, a new response comes each time.

"1000s of prompts have been attempted, not just by this plaintiff, but New York Times in the United States and still no instance of reproduction has been shown despite utmost attempts," he added.

Sibal also said that ChatGPT is trained never to regurgitate or reproduce the expression of any particular publicly available work.

He took the Court through the examples of reproduction cited by ANI to claim copyright infringement. Sibal said ANI had put prompts on ChatGPT in a manner to generate exact copy of the reportage.

"In the prompt itself, there is an extract of headline of article. This is more akin to plaintiff saying please give me my work," he submitted.

The Court today questioned whether excerpts from an exclusive interview can be reproduced by someone else.

"If he [journalist] has invested in resources to get this interview done, even though maybe your words are different or you're paraphrasing, question is, how do you get the right to reproduce his interview for which he hired reporters or invested money and resources," Justice Bansal asked.

Sibal answered that articles by other journalists are written about interviews with quotes from that exclusive interaction.

"When another journalist reads that [interview] and he narrates that in his own words in an article, it cannot be infringement. On the other hand, if he plagiarizes it such that [if] an ordinary person was reading the two ... will see it's a blatant act of piracy, then certainly ... but if he puts it in his own words, he can narrate whatever is in the article," he said.

Sibal also said that that an interviewer does not get ownership of the facts disclosed by an interviewee.

"If a person interviews someone and gets access to ideas for the first time, he doesn't get authorship or protection over those ideas, the facts that emanate but certainly his effort, his skill, putting down in his own expression, in narration of that gets protection to a limited extent," he submitted.

At this stage, advocate Sidhant Kumar for ANI submitted that such exception allowing narration or reportage of an exclusive interview would be applicable only in cases of media or journalists.

"That defense is not available to everyone," Kumar said. Sibal responded that the first stage in this case is whether there is infringement at all under Section 14 of Copyright Act.

Sibal also argued that copyright in the quotations vest with the speaker/ interviewee.

"Exclusive interview is even worse. Because exclusive interview will be actually a case where copyright in the quotations is owned by the speaker in the interview, not by the journalist. Because there is no originality by the journalist," he said.

The question of attribution also arose during the hearing. The Court asked whether a credit can be a justification for using the exact expression which a media organization has used.

However, Sibal responded that on ChatGPT, it would never be the exact expression that the media organization has used. At this stage, Kumar said similar attribution is not there in the examples cited by him

The Court then remarked that ChatGPT is a work in progress. Kumar agreed that technology is dynamic. Meanwhile, Sibal said attribution is also given in examples taken by ANI.

"There is no attempt to hide the source of information, facts," he further said.

Sibal will continue his arguments on April 29.