The Seattle Times and Newsday have sued Microsoft and OpenAI, accusing the companies of using their journalism without permission to build AI products that compete for their readers.
Filed on September 4 in federal court in New York, the lawsuit alleges the companies scraped hundreds of thousands of articles, including paywalled content, to train their models.
The complaint also directly challenges how AI search answers are assembled. It alleges that the companies continue to retrieve and copy the newspapers’ current reporting when answering questions about the news.
That process, known as retrieval-augmented generation, draws on material retrieved when a question is asked. It can use an article even if that article was never included in the model’s training data.
The newspapers say these answers leave readers with less reason to visit or subscribe. Their allegations cover ChatGPT, Copilot and Bing’s AI features.
One example in the filing concerns The Seattle Times’ Pulitzer-winning Boeing 737 MAX coverage. The publishers say ChatGPT reproduced 88 consecutive words verbatim after receiving a prompt containing the article’s headline and URL.
OpenAI told Reuters that its models are trained on publicly available data and that its approach is grounded in fair use, without addressing the lawsuit specifically. Microsoft said it was surprised by the suit and remained willing to discuss a resolution.
Both newspapers had previously joined an AI fellowship backed by the companies they are now suing. The Lenfest Institute’s $10 million program, announced in 2024, provided grants and technology credits to help newsrooms experiment with AI.
The publishers are seeking damages and an order requiring the destruction of copies of their work, along with training datasets and AI models incorporating it.
Start the conversation by posting the first comment