What happened
On September 4, 2026, The Seattle Times Company and Newsday LLC filed a lawsuit against OpenAI and Microsoft in the US District Court for the Southern District of New York. The complaint alleges the companies scraped the publishers' websites, including content behind paywalls, and incorporated the articles into the large-scale datasets used to train, fine-tune, and ground the large language models behind ChatGPT, Microsoft Copilot, and Bing's AI features. It further alleges the companies stripped or altered copyright management information, such as bylines and copyright notices, from the copied articles. The suit brings claims for copyright infringement, vicarious copyright infringement against Microsoft, removal of copyright management information under the Digital Millennium Copyright Act, and trademark dilution under the Lanham Act and Washington and New York state law. The complaint describes generative AI trained on journalism as "a snake eating its own tail" that could leave independent journalism "broken beyond repair," and calls AI products like ChatGPT and Copilot "rapacious consumers" of human-authored content rather than producers of new work. The publishers are not only seeking damages. They are asking the court for an order requiring the destruction of copies of their works, along with the training datasets and AI models that incorporate them, a far more severe remedy than a licensing fee or a settlement check. OpenAI has said publicly, without commenting on this suit specifically, that its models are trained on publicly available data and grounded in fair use. Microsoft said it is "surprised by the lawsuit" but open to discussing solutions, noting it has previously funded some of the Seattle Times' own journalism fellowships. Neither company has been found liable for anything; the case was just filed and will take time to move through the court. This is not the first suit of its kind. The New York Times sued OpenAI and Microsoft over similar allegations in 2023, and that case, along with several others, has been consolidated into a larger multidistrict litigation still working through discovery. What is new here is the specific remedy sought (destroying trained models and datasets, not just paying for past use) and the added trademark claim, which argues the reproduction and paraphrasing of the publishers' work also dilutes their brand.
Why it matters for business owners
A freshly filed complaint has no legal weight on its own. OpenAI disputes the claims, no judge has ruled on anything, and courts have never ordered a major AI company to destroy a trained model at this scale. The realistic range of outcomes, based on how similar cases have gone so far, runs from dismissal to a licensing settlement (the path Anthropic chose in a separate book-publishing case, paying $1.5 billion rather than litigate to a verdict) to a narrower ruling on specific infringing outputs. Destruction of a live, widely deployed model is the least likely outcome, and even the publishers' own filing acknowledges this is a demand, not something they expect to receive automatically. So nothing requires your business to change how it uses ChatGPT, Copilot, or Bing today. What is worth attention is not the lawsuit's odds of success, but what it reveals plainly: the same AI products many businesses now use daily for research, drafting, and customer-facing chat were built in part by reading and reproducing other people's content without asking, according to more than a dozen similar suits now working through US courts. And the specific harm the publishers describe (an AI answer that satisfies a reader without a visit to the original site) is not unique to news journalism. It is the same mechanism at work anywhere a business publishes something valuable online that an AI tool can read, summarize, and hand to someone else without sending that person back to the source.
What owners should not misunderstand
Do not read this as a sign that ChatGPT, Copilot, or Bing are about to disappear or get shut down. Even in the unlikely event the publishers eventually win a favorable ruling, courts overwhelmingly favor remedies like damages, licensing orders, or output-level restrictions over destroying a product millions of businesses and consumers depend on daily. The multidistrict litigation this case will likely be folded into has been running since 2023 without disrupting access to any of these tools. Do not assume this is only a newspaper industry problem, either. The publishers' complaint centers on a business model built on original content and a AI answer engine that reduces the reason to visit the source. Any business with a blog, a resource library, product guides, or original research on its own website depends on the same basic exchange: people find the content, and some of them convert into customers. If an AI tool increasingly answers the question directly, using content it read from your site or someone else's, without sending the reader through, that exchange quietly weakens regardless of who wins this specific lawsuit. And do not treat the trademark dilution claim as a minor add-on. It reflects a genuinely separate concern from copyright: that an AI product reproducing or closely mimicking a brand's original work, in a way readers cannot easily distinguish from the source, can wear down the value of the brand itself, not just the value of any one article. That is worth thinking about for any business whose reputation rests on being the trusted, original source for something specific.
The operational lesson
Most businesses that publish content online have never actually checked how AI tools are already treating that content. Two separate questions are worth answering, and this lawsuit is a useful prompt for both, independent of how the case itself resolves. First, exposure: ask a few AI chat tools direct questions your content is written to answer, and see whether they answer from your material, summarize it accurately, cite it, or reproduce it closely enough that a reader would have no reason to click through. This takes twenty minutes and tells you more about your actual AI exposure than any news story will. Second, dependency: if your business relies on ChatGPT, Copilot, Bing AI, or a similar tool for a workflow that touches customer-facing content, research, or drafting, this suit is one more entry in a growing list of unresolved legal questions about how these tools were built. That is not a reason to stop using them today. It is a reason to avoid designing a workflow so tightly around one vendor's specific tool that a future licensing change, price increase, or feature restriction (any of which is more likely than a court-ordered shutdown) would be disruptive to unwind.
What a serious business should do next
Do not switch AI vendors, pause an AI project, or make any legal assumptions based on this lawsuit. It is an early-stage complaint with a disputed set of facts and a resolution that is years away. Do spend twenty minutes checking how a few mainstream AI tools answer questions your own website content is built to answer, so you know your actual exposure instead of guessing. Do keep a record, even an informal one, of your original content and when it was published, the same instinct that leads a large publisher to register its work and track its readership. If your content becomes a meaningful part of how customers find you, knowing what you have and when you made it is useful regardless of how AI copyright law develops. Do track this case and the broader multidistrict litigation loosely, not for the legal outcome itself, but for what it signals about how AI vendors source their systems and how much scrutiny that sourcing is now under. A ruling that meaningfully changes how AI companies can train on or reproduce copyrighted content would eventually affect what these tools can do and how they are priced, and that is worth knowing before it happens, not after.
The Atlacis view
It is easy to read a story about newspapers suing AI companies and conclude it has nothing to do with a medium-size business. It does. The mechanism the publishers describe, an AI tool reading original content and giving readers an answer that removes their reason to visit the source, plays out at a much smaller scale for any business that has invested in content to attract and convert customers. Atlacis helps business owners understand where their own content, workflows, and customer data actually sit in relation to the AI tools they use, what depends on a single vendor's product decisions, and what to check before assuming an AI tool is representing your business accurately, independent of how any single lawsuit against a major AI vendor turns out.
The short version
- On September 4, 2026, The Seattle Times and Newsday sued OpenAI and Microsoft, alleging their journalism was scraped, including paywalled content, and used without payment to train ChatGPT, Copilot, and Bing's AI features.
- The suit brings copyright infringement, DMCA copyright-management-information removal, and Lanham Act trademark dilution claims, and it asks the court to order the destruction of the AI models and training data built on the publishers' work, not just damages.
- This is a freshly filed complaint, not a ruling. OpenAI disputes the claims, and courts have never ordered a major AI model destroyed; a licensing settlement or a narrower ruling is a far more likely outcome than a shutdown.
- Nothing requires any business to change how it uses ChatGPT, Copilot, or Bing today because of this filing.
- The mechanism at the center of the suit, an AI answer satisfying a reader without a visit to the original source, applies to any business that publishes original content on its own website, not only to news publishers.
- The useful response is not a legal one. Spend twenty minutes checking how AI tools already answer questions your content is built to answer, and avoid building a workflow so tightly around one AI vendor that a future licensing or pricing change would be disruptive to unwind.
Where ATLACIS can help
Sources
- The Seattle Times Company and Newsday LLC v. OpenAI, Inc. et al., Complaint (US District Court for the Southern District of New York, filed September 4, 2026)
- Reuters via Honolulu Star-Advertiser: Seattle Times, Newsday sue OpenAI, Microsoft, alleging copyright infringement (September 5, 2026)
- TechCrunch: Seattle Times and Newsday are the latest publications to sue OpenAI and Microsoft (Anthony Ha, September 5, 2026)