The Seattle Times Newsday OpenAI lawsuit has put an unusually broad remedy before a federal court: the two publishers want OpenAI and Microsoft to destroy not only allegedly unauthorized copies of their journalism, but also training datasets and AI models said to contain it. The companies were sued in the U.S. District Court for the Southern District of New York on Sept. 4, according to Reuters.
The newspapers allege that OpenAI and Microsoft copied their reporting without permission to train and operate AI products. Their claims have not been decided by a court.
According to Reuters, the complaint says the companies scraped the publishers' websites, including articles behind paywalls, then put that material into datasets used for ChatGPT, Microsoft Copilot and Bing's AI features.
What do the Seattle Times and Newsday allege OpenAI and Microsoft did?
The publishers contend that the products can return passages from their reporting or closely rework articles in response to user prompts. They say those answers can give readers information without a visit to the original publisher, reducing traffic and subscription sales.
That allegation reaches beyond the familiar question of whether a chatbot output is too close to a particular article. The newspapers are also challenging the alleged upstream use of their work in the data and models behind the products.
Seattle Times President and CEO Alan Fisco told employees that the newspaper spends millions of dollars each year producing journalism and needed to protect that work from use without consent or compensation, Reuters reported.
The publishers seek relief aimed at models, not only articles
The Seattle Times and Newsday are seeking an order requiring the destruction of copies of their work, plus training datasets and AI models that incorporate it, according to Reuters. That request, if granted, would target material used to build or run the systems, rather than merely requiring removal of identified articles or outputs.
No ruling has established that the companies used the newspapers' material as alleged, or that a court will grant the requested remedy.
OpenAI cites fair use; Microsoft says it will discuss solutions
OpenAI did not address the case specifically. A spokesperson said the company trains its models on publicly available data and considers its practices grounded in fair use, Reuters reported.
Microsoft said it was surprised by the filing. In a statement reported by Reuters, the company said it recognized the importance of local journalism and was open to discussing solutions for this type of dispute.
The case joins a widening set of copyright fights over generative AI training. Reuters said The New York Times' 2023 lawsuit against OpenAI and Microsoft remains pending, and described the new action as one of dozens brought by copyright holders against AI companies including OpenAI, Anthropic and Meta.
This story draws on original reporting from The Verge.
