ETtech Explainer: Legal tussle between OpenAI and NYT

OpenAI's investor Microsoft and its co-defendant in the New York Times (NYT) lawsuit on Monday sought to dismiss parts of NYT's plea on the grounds that the news organisation did not show any “actual harm” or meaningful diversion of revenue from its website due to use of artificial intelligence (AI) models like ChatGPT.

The NYT had sued the two tech firms in December 2023, alleging that they were trying to create a substitute of the newspaper and even trying to “free-ride on the Times’s massive investment in its journalism”.

ET explains the legal battle and the course of events so far.


What is the lawsuit about?

In its petition filed in the US District Court for the Southern District of New York, the NYT said the organisation is losing revenue because of AI tools which are diverting readers from its websites and even producing responses which are near verbatim excerpts from its articles.

Also read | ETtech Explainer: OpenAI’s response to NYT’s copyright lawsuit

The NYT also produced evidence of prompts giving such responses to the court.

Following the news organisation’s motion, several authors, computer programmers and musicians moved court against big tech companies, including Microsoft, OpenAI, Meta and GitHub, for training their models on copyrighted information without fairly compensating for it.

How did the defendants respond?

OpenAI came down heavily on the NYT, accusing it of “hacking” ChatGPT and making thousands of attempts to generate “the highly anomalous results".

“In the ordinary course, one cannot use ChatGPT to serve up Times articles at will,” OpenAI said.

To this, the news organisation responded that what OpenAI bizarrely mischaracterised as “hacking” is simply using OpenAI’s products to look for evidence.

OpenAI also said that the news organisation “cannot prevent AI models from acquiring knowledge about facts, any more than another news organisation can prevent the Times itself from re-reporting stories it had no role in investigating”.

Separately, in Monday’s motion, Microsoft argued that large language models did not supplant the market for news articles and other materials they were trained on.

ChatGPT “is not in any way a substitute for a subscription to The New York Times”, it said.

Why are AI companies facing such lawsuits?

AI companies have built large language models by training them on enormous amounts of public data, some of which is likely copyrighted. They claimed that such material can be used for training because it is public and they are not producing a tool to replicate the information in its entirety.

The courts have dismissed some cases on lack of evidence that content produced by generative AI tools resembles copyrighted work. But, they have not yet addressed the key question of whether AI training qualifies as fair use under copyright law.