'New York Times' sues over AI using its articles without permission
ARI SHAPIRO, HOST:
The New York Times is suing the creator of ChatGPT. The Times claims OpenAI developed the chatbot by using articles published in the newspaper without permission. It's the latest in a growing number of copyright infringement suits OpenAI is facing. And to help us make sense of this, we are joined by NPR tech correspondent Bobby Allyn. Hey, Bobby.
BOBBY ALLYN, BYLINE: Hey, Ari.
SHAPIRO: What exactly does The New York Times claim in this suit?
ALLYN: Essentially that OpenAI's popularity and profits depend in part on its copying of millions of New York Times articles. Now, to back up for a moment, the way an AI chatbot like ChatGPT works is it hoovers up vast amounts of material from the internet. That chatbot is then trained on this material. And when someone like you or I asks it a question, all of its answers are informed by the text. And some of that text is copyrighted. That means it is not for free.
Now, I reported back in August, Ari, that the Times threatened to sue OpenAI after talks about a potential licensing deal fell apart. And today, The Times officially gave up and decided to drag OpenAI to court over this. The Times is also suing Microsoft, which is OpenAI's biggest financial backer.
SHAPIRO: And what have OpenAI and Microsoft said in response to the suit?
ALLYN: So far, the companies haven't responded directly. But in the past, OpenAI has defended its mass scraping of the internet by citing something called fair use doctrine, which is a legal theory that basically says, in certain circumstances, like for research, teaching, criticism, parody, it's OK to use copyrighted material without permission. OpenAI says it's relying on huge swaths of the internet under fair use law.
Now, the Times, of course, is not buying that. In its lawsuit, the paper says ChatGPT is actually stealing its audience away from the paper - basically, that ChatGPT has become a competitor of the Times by serving up portions of its news articles with no payment or benefit to the paper.
SHAPIRO: The New York Times is obviously a big player in the news industry. Could this lawsuit affect all of digital publishing?
ALLYN: You know, it could. The question of whether AI companies have followed copyright law in building these super powerful chatbots is unresolved, and many other publishers have the same concerns. And it really strikes a nerve for many media organizations that have become increasingly leery of Silicon Valley news outlets - as we know, have long been reliant on search engines and social media. You know, that's how we reach digital audiences. But, you know, it was a promising partnership that really soured fast. You know, tech companies have moved away from the news just as they're pocketing lots and lots of digital advertising dollars, and media companies like the Times want to avoid repeating history with the rise of AI.
SHAPIRO: And just briefly, what are some possible outcomes of this suit?
ALLYN: Yeah, you know, Ari, this suit can be really damaging for OpenAI. The worst possible outcome for them is that a judge orders them to destroy the dataset that ChatGPT is built on. And I've talked to many legal and technical experts who say, I don't know how ChatGPT would function at all without this dataset. Now, it's a pretty extreme outcome, but it is within the realm of possibility, so we will see how this lawsuit shakes out.
SHAPIRO: NPR's Bobby Allyn. Thank you.
ALLYN: Thanks, Ari.
(SOUNDBITE OF SLVR SONG, "BACK N FORTH") Transcript provided by NPR, Copyright NPR.
NPR transcripts are created on a rush deadline by an NPR contractor. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.