A 38-page complaint filed September 4, 2026 in the United States District Court for the Southern District of New York, No. 1:26-cv-07644, by The Seattle Times Company and Newsday LLC through Klaris Law PLLC (Lacy H. Koonce, III and Brendan T. Kehoe). A jury trial is demanded.
The named defendants are nine OpenAI entities — OpenAI, Inc.; OpenAI GP, LLC; OpenAI, LLC; OpenAI OpCo, LLC; OpenAI Global LLC; OAI Corporation; OpenAI Holdings, LLC; OpenAI Foundation; and OpenAI Group PBC — together with Microsoft Corporation. The entity list tracks the post-restructuring corporate structure and is pleaded so that the vicarious count can be directed at particular entities.
The claims
Seven counts are pleaded. The two trademark-dilution theories under state law and the federal dilution count distinguish this filing from the copyright-only publisher suits consolidated before the same court.
| Count | Claim | Statute |
|---|---|---|
| I | Copyright infringement | 17 U.S.C. § 501 |
| II | Vicarious copyright infringement | — (against Microsoft, OpenAI Inc., OpenAI GP, OpenAI LP, OAI Corporation, OpenAI Holdings LLC and OpenAI Global LLC) |
| III | Removal of copyright management information | 17 U.S.C. § 1202(b)(1) |
| IV | Distribution of works with CMI removed | 17 U.S.C. § 1202(b)(3) |
| V | Trademark dilution | 15 U.S.C. § 1125(c) |
| VI | Trademark dilution (Seattle Times only) | RCW 19.77.160 (Washington) |
| VII | Trademark dilution (Newsday only) | N.Y. Gen. Bus. L. § 360-L |
The dilution theory
The dilution counts rest on registered word marks: THE SEATTLE TIMES, U.S. Registration Nos. 5,738,989, 5,738,990, 5,738,991 and 5,738,992, each registered April 30, 2019 and pleaded as incontestable under 15 U.S.C. § 1065 following acceptance of combined Section 8 and 15 declarations on August 22 and 25, 2025; and NEWSDAY, U.S. Registration No. 2,047,787, registered March 25, 1997, renewed and pleaded as incontestable. First use in commerce is pleaded as 1896 and 1940 respectively.
Two dilution mechanisms are pleaded separately. Blurring under § 1125(c)(2)(B): the defendants' products "reproduce, output, and associate Plaintiffs' Marks with content generated by Defendants' own models, including content that is not Plaintiffs' authentic journalism," which the complaint says whittles away the source-identifying association between the marks and the plaintiffs' journalism. Tarnishment under § 1125(c)(2)(C): the products "hallucinate" content misattributed to the plaintiffs, or attribute the plaintiffs' actual content to another source, associating the marks with material the plaintiffs "did not create, did not review, and would not publish."
Willfulness is pleaded on the basis that the dilutive uses continued after notice from the publicly reported publisher litigation.
The copying allegations
The complaint pleads, on information and belief, that the defendants obtained copies by scraping the plaintiffs' websites with automated bots and bypassing their paywalls; incorporated those copies into large-scale datasets including those derived from WebText, WebText2, Common Crawl and Microsoft's Bing search index; and used them to train, fine-tune and ground the models behind ChatGPT, Copilot and Bing Chat. It further pleads copying for retrieval-augmented generation, both through web crawls building indices and through direct scrapes of the plaintiffs' sites in response to user queries about current events.
On regurgitation, the complaint alleges that models reproduced substantial passages "even when supplied with nothing more than the article's headline, publication date, and URL," which it offers as evidence that the text was recalled from training rather than from user-supplied input. The specific example given at paragraph 66 is an article from the Seattle Times' Pulitzer-winning series on the 2019 Boeing 737 MAX crash and the ensuing federal investigation, from which the complaint says the model "regurgitated 88 consecutive words verbatim" in response to a prompt containing the headline and URL. Paragraph 67 sets two Newsday passages beside model output in a side-by-side table.
Fair use, licensing and harm
The complaint anticipates the fair-use defence directly, pleading that there is "nothing transformative" about copying journalism to build products that compete for the same readers, advertisers and subscription dollars, and that the outputs "substitute for a visit to newsday.com or seattletimes.com."
It pleads that OpenAI has licensed content from more than a dozen news and media organisations — naming, on information and belief, the Associated Press, News Corp, Axios, Axel Springer, The Atlantic, the Financial Times, Dotdash Meredith and Vox Media — and that the disclosed terms of three of those agreements reflect payments of more than $300 million, offered as evidence both that a licence is required and that a market exists.
On harm, the complaint cites industry data for December 2024 to December 2025 showing search referral traffic to mid-sized regional and metro daily publishers falling by approximately 47%, against approximately 22% for larger national publishers, attributing the difference to mid-sized publishers' heavier historical dependence on incidental search traffic.
Relief sought
The prayer asks for statutory damages or actual damages and profits under 17 U.S.C. § 504, at the plaintiffs' election; statutory or actual damages for each § 1202 violation; a permanent injunction; and, under 17 U.S.C. § 503, "the impoundment and/or destruction … of all copies of Plaintiffs' works, and all LLMs and training datasets incorporating Plaintiffs' works or derivatives thereof, in Defendants' possession, custody, or control." On the dilution counts it asks for an injunction, an accounting of profits, damages, and — "in light of Defendants' bad faith and willful conduct" — treble damages and attorneys' fees under 15 U.S.C. §§ 1117(a) and 1125(c)(5), RCW 19.77.160 and N.Y. Gen. Bus. L. § 360-L.
Jurisdiction and related cases
Personal jurisdiction is pleaded on the defendants' New York presence: OpenAI's offices at the Puck Building, 295 Lafayette Street, and approximately 450 New York City employees as of January 2025; Microsoft's offices at 11 Times Square, 122 Fifth Avenue and 300 Lafayette Street, and LinkedIn's roughly 500,000 square feet at the Empire State Building. The complaint also pleads that the defendants have consented to jurisdiction and venue in the district in The New York Times Company v. Microsoft Corp., No. 23-cv-11195; Raw Story Media, Inc. v. OpenAI, Inc., No. 24-cv-01514; The Intercept Media, Inc. v. OpenAI, Inc., No. 24-cv-01515; and Daily News, LP v. Microsoft Corp., No. 24-cv-03285 — all consolidated in In re OpenAI, Inc. Copyright Infringement Litig., No. 25-md-03143 (NYT v. Microsoft, OpenAI et al.).
Provenance
Exhibits 1 through 4 — the schedules of registered works and the trademark registration certificates — are not part of the filed document captured here.
Relationships
- related: The Seattle Times Co. and Newsday LLC v. OpenAI and Microsoft — the case-tracking page.
- related: NYT v. Microsoft, OpenAI et al. — the consolidated multidistrict proceeding whose jurisdictional record the complaint invokes.
- related: OpenAI; Microsoft — the defendants.
- related: Media, Journalism & Entertainment — AI Deployment; AI Copyright — the sector and the doctrine.
- supports: AI Copyright — extends the publisher litigation from copyright and DMCA theories to trademark dilution.