Seattle Times and Newsday sue OpenAI and Microsoft over AI training data

The Seattle Times and Newsday have sued OpenAI and Microsoft for copyright infringement, meaning the use of protected work without permission. The newspapers say OpenAI took their articles as training data, the text fed into a model to shape what it can do, and never asked or paid for it. They also say OpenAI’s chatbots often reproduce passages from their reporting word for word when users ask questions. Those are two separate objections: how the models were trained, and what they produce once released.

The two publishers want a remedy that goes beyond money. They are asking the court to make OpenAI and Microsoft destroy every copy of their works the companies hold, the training datasets that contain those works, and the AI models built using them. In practice that would mean deleting systems that already run inside products people use every day.

Microsoft is named as a defendant because its Copilot assistant is built on OpenAI’s technology. The suit treats the two companies as jointly responsible for how the models were trained and for what they produce in response to user queries.

The case adds to a growing group of similar claims against OpenAI. The New York Times, Ziff Davis, Merriam-Webster and Encyclopedia Britannica have each taken the company to court over training data. Separately, close to 400 local newspapers recently sued both OpenAI and Microsoft, arguing that AI chatbots answer readers’ questions well enough that fewer people click through to news sites, which cuts into subscription revenue. The common thread across these suits is the claim that building a paid AI product on copyrighted articles, without a license, breaks the law.

OpenAI and Microsoft did not immediately respond to The Verge’s request for comment. The report does not give the date the suit was filed, the court hearing it, the detailed legal claims, or a figure for the damages sought. It also does not say whether the Seattle Times and Newsday want their case joined to the New York Times lawsuit or kept separate.

What to watch next is how the court handles the request to destroy models and training datasets, and whether OpenAI or Microsoft respond in public to the claim that their chatbots repeat article text. The other open question is whether this complaint is combined with the other publisher cases against the two companies or moves ahead on its own.