The Seattle Times and Newsday have filed lawsuits against OpenAI and Microsoft, alleging that paywalled journalism was used without permission to train and operate the companies' artificial intelligence models. The filings, reported on 8 September 2026, place the two regional publishers alongside The New York Times and Ziff Davis, which already have separate copyright actions running against the same defendants. What distinguishes the new cases is that both plaintiffs previously took part in a grant programme funded by the companies they are now suing.
What the complaints allege
According to the reporting, the two publishers allege that their paywalled content was used to train and operate OpenAI's and Microsoft's models. They are asking the court for remedies that include the destruction of datasets and models trained on their work. These are allegations contained in court filings. They have not been proven, and the fair use questions at the centre of the dispute remain unsettled in the ongoing cases.
The earlier grant programme
Both newsrooms were recipients of money from a $10 million initiative called the AI Collaborative and Fellowship, funded by OpenAI and Microsoft and created in collaboration with the Lenfest Institute for Journalism. Announced in October 2024, the programme paid for participating organisations to hire two-year AI fellows and was framed around business sustainability and innovation. The Seattle Times and Newsday used it to build tools for advertising, data summarisation and aggregation. The Minnesota Star Tribune and The Philadelphia Inquirer also took part. The existence of the grants is not in dispute. Whether accepting grant funding has any bearing on the copyright claims is a separate question that the litigation has not yet answered.
Washington enters the New York Times case
The legal environment around the new suits is complicated by federal intervention in the older one. In response to the complaint filed by The New York Times on 27 December 2023, the Trump administration submitted a Statement of Interest to the US District Court for the Southern District of New York. The filing argues that the newspaper "seeks to narrow fair-use doctrine to exclude the training of OpenAI's large language models", a result it calls inconsistent with basic copyright principles and harmful to "the Progress of Science and useful Arts". It also contends that restricting AI training would hurt smaller publications, and states that "the United States has a strong interest in this Court rejecting any argument that training LLMs on copyrighted texts violates copyright law".
The New York Times rejected that position. Its spokesperson, Graham James, said that permitting AI companies to use copyrighted material without permission would undermine the sustainability of human-created content. Although the statement was filed in the Times case, its reasoning could be invoked in the other pending suits, including those brought by The Seattle Times and Newsday.
Responses from the defendants
Microsoft said it was surprised by the Seattle filing but would cooperate to resolve any issues. Both companies deny wrongdoing. OpenAI maintains that the material used to train its models adheres to fair use law. None of the cases has produced a ruling on that question.
What is established and what is merely claimed
Established: The Seattle Times and Newsday have filed lawsuits against OpenAI and Microsoft. Both previously received grants from the $10 million AI Collaborative and Fellowship funded by the defendants and run with the Lenfest Institute, announced in October 2024. The Trump administration has filed a Statement of Interest in the separate New York Times case, urging the court to treat AI training on copyrighted text as lawful. The New York Times and Ziff Davis suits remain ongoing. OpenAI and Microsoft deny wrongdoing.
Merely claimed: that paywalled content from the two publishers was in fact used to train or operate the models, which is an allegation in the complaints. That such use, if it occurred, constitutes copyright infringement, which the plaintiffs assert and the defendants dispute. That training on copyrighted text is protected fair use, which is the position of OpenAI and of the administration's filing, not a holding by any court in these cases. And that the appropriate remedy includes destroying datasets and models, which is what the plaintiffs have requested, not what any court has ordered.