Another day, another lawsuit against the model builders.
On September 4, 2026, The Seattle Times Co. and Newsday filed a federal copyright and trademark complaint in the U.S. District Court for the Southern District of New York, joining a growing line of newsrooms taking OpenAI and Microsoft to court. The claim is straightforward and familiar: the companies used their journalism without permission to train AI systems, and the training itself harms the business of producing news.
From a technical standpoint, this is the most consequential fight happening in AI right now. Not because the outcome will be clean — it won’t be — but because the lawsuit forces a hard question that the industry has been dancing around for years: what exactly happens to a machine’s knowledge when you remove the text it was built from?
The Training Set as a Production Line
Let me get technical here. When I say a model “learns” from news articles, I’m not being metaphorical about reading. A large language model isthe model. And the people who made the record are due a say in what it becomes.
🕒 Published: