Eight American newspapers, including Chicago Tribunefiled a lawsuit Tuesday against OpenAI, creator of ChatGPT, and Microsoft, its main investor, accusing the companies of violating their copyrights to create the technology used in artificial intelligence (AI) assistants.
• Read also: Agreement between OpenAI and the “Financial Times” to train AI systems in “journalism”
• Read also: Complaint in Austria against ChatGPT, “unable to correct” its errors
• Read also: United States: creation of an advisory committee on the “safe and secure” use of AI
“This lawsuit stems from the fact that (the companies) stole millions of publishers’ copyrighted articles, without permission or payment, in order to market their generative artificial intelligence products, including ChatGPT and Copilot (from Microsoft),” says the complaint filed in a court in New York.
The plaintiffs – New York Daily News, Chicago Tribune, Orlando Sentinel, Sun Sentinel of Florida, San Jose Mercury News, Denver Post, Orange County Register And St. Paul Pioneer Press – belong to Alden Global Capital, a Florida-based hedge fund that created the second largest American newspaper group behind Gannett, owner of USA Today.
According to their lawyers, “this trial will demonstrate that the defendants must both obtain the consent of publishers to use their content and compensate them fairly for this use.”
They also criticize AI assistants for offering excerpts from articles and attributing misleading or inaccurate information to publications in some cases.
Many authors, artists and information sites accuse OpenAI and its rivals of copyright infringement in the race for generative AI (production of texts, images, etc., upon simple request in everyday language), which requires mountains of data. At the end of December, the New York Times launched lawsuits against OpenAI and Microsoft.
OpenAI defends itself
Asked by AFP on Tuesday, OpenAI did not comment on the newspapers’ allegations, but said it “takes great care” in its products and its process of training generative AI models “to support organizations in press”.
The Californian start-up also highlighted “constructive partnerships and conversations with numerous media outlets around the world to explore opportunities, discuss concerns and provide solutions”.
It refers to content licensing agreements concluded with media outlets — including the press agency APthe German group Axel Springerthe French daily The world and the Spanish conglomerate Prisa Mediaand since Monday, the British daily Financial Times.
In the case of New York TimesOpenAI vigorously defended itself, arguing that using publicly available data, including news articles, for general model training did not constitute copyright infringement.
The start-up also accused the American daily of having “hacked” ChatGPT, to produce “highly abnormal results”.
Microsoft declined to comment on the complaints filed.
