The Seattle Times and Newsday have sued OpenAI and Microsoft, alleging their journalism was used without permission to train the large language models behind ChatGPT and Copilot. The two regional dailies join a lengthening line of US publishers pursuing the same claim against the same two defendants. Neither the filings' damages figures nor a case schedule have been made public in the initial reporting.
The suits accuse OpenAI and Microsoft of copying and ingesting protected articles at scale to build commercial AI products that now compete with the publishers for reader attention and ad dollars. That framing mirrors the theory of the case in the earlier New York Times lawsuit against the same two defendants, which remains the anchor litigation for the industry.
OpenAI has not yet responded publicly to the Seattle Times and Newsday filings. Microsoft has not yet responded either. Both companies have consistently argued in prior cases that training on publicly available text qualifies as fair use under US copyright law, and that model outputs are transformative rather than substitutes for the original works.
Key facts
- 01The Seattle Times and Newsday filed copyright infringement suits against OpenAI and Microsoft over the use of their journalism in AI training.
- 02The filings extend the list of US news publishers suing the two companies, which already includes The New York Times and a coalition of daily papers.
- 03OpenAI and Microsoft have publicly maintained that training on publicly available text falls within fair use.
- 04Microsoft recently disclosed that Copilot rarely reproduced New York Times articles across 8.2 million analyzed chat logs, a figure central to its defense in that parallel case.
The Seattle Times is the largest daily newspaper in the Pacific Northwest and has run its own subscription-based digital operation for years. Newsday, based on Long Island, is one of the largest suburban dailies in the country. Both fit the profile of regional publishers that have watched their digital ad market shrink while AI products increasingly answer reader questions directly, without the click that pays their reporters.
The wave of publisher litigation has bifurcated. Some outlets, including the Financial Times, Axel Springer, News Corp, and The Atlantic, have signed licensing deals with OpenAI that pay for training access and, in some cases, real-time content use inside ChatGPT. Others, led by The New York Times, have chosen to sue rather than settle, betting that a court win or a large settlement will set a higher price for the industry than a negotiated deal today.
The Seattle Times and Newsday filings put them in the litigation camp. Whether that reflects a strategic bet on damages, dissatisfaction with the terms on offer, or simply the fact that no offer has been made to these specific outlets is not addressed in the initial coverage. Regional papers have generally received less attention from the AI labs' licensing teams than the national brands, a gap that has fed frustration inside the trade.
Microsoft is a co-defendant in almost all of these cases because it hosts and distributes OpenAI's models through Azure and embeds them in Copilot across Windows, Office, and Bing. The company recently disclosed, in the parallel New York Times case, that Copilot reproduced Times articles only rarely across 8.2 million analyzed chat sessions, a data point it is using to argue that any infringement is de minimis rather than systemic. That defense will presumably extend to the new suits.
The plaintiffs' theory does not depend on the models spitting out full articles verbatim. It rests on the copying that happens during training itself, before any user ever types a prompt. Courts have not yet issued a merits ruling on that theory in an AI-training case at the appellate level, which is why every new filing raises the stakes on the eventual decisions.
Discovery in the New York Times case has already surfaced granular data about how OpenAI assembles training corpora and how often Copilot's outputs overlap with copyrighted text. Filings from Seattle Times and Newsday would likely draw on the same evidentiary trail, which could accelerate their proceedings if the cases are consolidated or coordinated. That is a decision for the assigned judges, not the parties.
For OpenAI and Microsoft, the tactical calculus does not change much with two more plaintiffs. The legal exposure was already existential in the sense that a bad ruling on training-data fair use would reshape the economics of the entire model industry, and adding two regional papers to the caption does not move that needle. What it does change is the political pressure — every additional lawsuit makes a legislative response, or a large industrywide settlement, marginally more likely.
For the newspapers, the suits are as much a negotiating posture as a legal strategy. A publisher that files gets a seat at the table for any future settlement framework in a way that a publisher that stays silent does not. Given how few regional dailies have signed AI licensing deals to date, expect more of these filings before the New York Times case reaches a verdict.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
