After Meta, it seems now it is the turn of OpenAI’s to get into trouble with Indian authorities and the law of the land. But unlike Meta, to make the matters worse for itself, OpenAI has a who’s who of Indian media on its wrong side who have joined and threw weight behind the original petitioner book publishers association who have accused OpenAI of using their content without their permission to train its AI bot ChatGPT.
But instead of striking a deal with these content creators – book publishers and news organisations – OpenAI is defiant saying the content is “freely” available and it is using it in the interest of public.
To make matters worse for itself, it is needling the Indian courts by questioning their jurisdiction, saying Indian law does not apply on as OpenAI does not have any business presence in India. This is an argument which courts have thrown out of window decades ago when cross-border digital paratroopers indulging in activities in Indian digital space while keeping their physical presence (servers) in other countries. Nothing can annoy courts more thoroughly than this. By that argument hacking can never be a crime because hackers do not have an office here.
Argument put forward by OpenAI in Delhi High Court that the content it was lifting was freely and openly available, is just specious. It is important to understand that the “content” it says is “freely available” is actually a “product” of petitioners (the content creators, book publishers and media houses). And like any other product, there is an investment that has gone into it, from salaries of journalists, writers. designers, and a batteriy of editors in the backend to operations and infrastructure. There are journalists who have burned midnight oil to produce that content that OpenAI is saying “freely” available. Not to mention the huge everyday running expenditures. So it has not come free. It is like telling a brand that your product was “openly” available, so I picked it up and put it up in my showcase.
Even from the pragmatic angle, how helpful such a confrontation with media will be, when OpenAI which is in race to grab a pie in world’s largest digital market with highest number smartphone users, it will have think for itself. Unlike other small countries, Indian media is still a veritable force that can ramp up campaigns that has made even powerful governments to reverse their policies and businesses to wind up.
Then, there is an altogether different aspect of this challenge. AI is not a perfect thing and is known to make mistakes. Brand reputation of original news sources also gets hit because of the false or misattribution in reports generated by AI. So the argument that it quoting the sources and in a way promoting those brands is not going down anyone’s throat.
Petitioners in this matter which include media heavyweights such as Hindustan Times, Indian Express, NDTV, news agency ANI and book publishers such as Bloomsbury, Penguin Random House, and Cambridge University Press have demanded that OpenAI cease using their copyrighted material or provide compensation. But OpenAI has rejected the demand and has asserted that it is going by the fair use policy of publicly available data.
Challenging the Indian courts’ jurisdiction, neither compensating nor making some content sharing business arrangement, is not going to help the godfather of ChatGPT.
It will be mistake on part of ChatGPT to treat its win in legal battle in November 2024 with American news outlets Raw Story and AlterNet as some kind of precedent. The U.S. District Judge Colleen McMahon dismissed the plea against OpenAI of using their articles without permission to train its language models. But the dismissal was based on failure of news outlets to provide evidence of harm and was not on the merit of their challenge. The judge allowed the news outlets to amend their plaint and come back again.
Almost all cases around the world against OpenAI are related to an unauthorized use of copyrighted materials to train AI models. These cases highlight the ongoing global tension between AI development and the protection of intellectual property rights. As AI models increasingly rely on vast datasets, often sourced from existing copyrighted materials, the legal frameworks governing such use are being actively challenged by all the content creators. The way to resolution does not go through the corridor of courts but through the company’s own policy cabinets.