MARY LOUISE KELLY, HOST:
The world's leading artificial intelligence company is under legal assault. OpenAI, maker of ChatGPT, was most recently sued by The New York Times for copyright infringement, but it's also facing suits from book authors, artists, music labels and others. NPR tech correspondent Bobby Allyn is here to explain. Hey there.
BOBBY ALLYN, BYLINE: Hey, Mary Louise.
KELLY: What are all these lawsuits against the owner of ChatGPT - what are they all about?
ALLYN: Yeah. You know, they all come down to data and how this data is being used. OpenAI, in developing popular tools like ChatGPT and the image creator Dall-E, crawled the entire internet for text and images and other material. And from that enormous bucket of data, these AI tools learn and are able to create something new.
The big legal question now is, did the company break the law by scraping up parts of the internet that contained copyrighted material? And if they did this, you know, Mary Louise, with a licensing agreement, then it would be fine. But instead, OpenAI did it without permission and without payment. I talked to Ed Klaris about this. He's a former general counsel at The New Yorker, and he's an expert in intellectual property law.
ED KLARIS: You can't just go and steal huge archives and then say, it's too hard for me to get rid of it. It seems kind of contrary to the rule of law.
KELLY: Contrary to the rule of law. OK, what is OpenAI saying to defend itself?
ALLYN: Yeah. Executives there have long pointed to something in the law known as fair use doctrine, and it says that if copyrighted material is quoted or used in some way in news reporting, research, criticism, you don't need to ask for permission or pay anyone.
But, you know, there are two tests of fair use that are going to be really important here, and the first is, are you transforming the original work into something new, and does that new thing you're creating compete with the original? Now, OpenAI says ChatGPT answers are something new and that it is not competing with The New York Times. IP lawyer Klaris says OpenAI may have a hard time proving to courts that ChatGPT's responses that include New York Times article snippets are truly transformative.
KLARIS: There needs to be some commentary, some additional perspective. A person using the original content on a fair use basis would have to add something new.
KELLY: Bobby, I'm thinking ChatGPT, OpenAI - obviously, not the only players in the AI field. What effect could these court cases have for the whole industry?
ALLYN: Yeah, you know, The New York Times, in its suit, claims that OpenAI took millions of its articles illegally. If a judge sides with the Times, OpenAI could be on the hook for billions of dollars. The Times also asks the court to order OpenAI to destroy any of its models that use work from the Times. That could potentially upend the company and upend the entire AI industry in a really big way because this is how all of these AI companies build their underlying models. Other AI companies are, of course, being sued for copyright infringement, and those suing include Getty Images over licensed photos, Universal Music over copyrighted song lyrics. Authors and comedians have taken AI companies to courts.
You know, Mary Louise, on the other side of this, though, there's publishers, like the Associated Press and Axel Springer, which publishes Politico, which are deciding to take another tack, which is striking a licensing deal with OpenAI. But for those, you know, deciding to not play nice with AI companies, I think it's going to take some time before the courts and likely the U.S. Supreme Court really clarifies here what's allowed and what's not.
KELLY: Thank you, Bobby.
ALLYN: Thanks, Mary Louise.
KELLY: NPR's Bobby Allyn. Transcript provided by NPR, Copyright NPR.