Seattle Times and Newsday sue on 4 Sep

On 4 September 2026, The Seattle Times Company and Newsday LLC filed a 38-page complaint in the Southern District of New York against OpenAI entities and Microsoft Corporation. The ECF header reads Case 1:26-cv-07644, Document 1, Filed 09/04/26. Counsel of record are Lacy H. ("Lance") Koonce, III and Brendan T. Kehoe of Klaris Law PLLC. The caption demands a jury trial.

Plaintiffs are The Seattle Times Company (Seattle, Washington) and Newsday LLC (Long Island / New York City). Named OpenAI defendants include OpenAI, Inc.; OpenAI GP, LLC; OpenAI, LLC; OpenAI OpCo, LLC; OpenAI Global LLC; OAI Corporation; OpenAI Holdings, LLC; OpenAI Foundation; and OpenAI Group PBC, plus Microsoft Corporation.

The complaint states defendants have consented to jurisdiction and venue in related publisher suits consolidated in In re OpenAI, Inc. Copyright Infringement Litig., No. 25-md-3143 (S.D.N.Y.). This Document 1 is a new complaint. All factual claims below are plaintiff allegations unless labeled otherwise.

Seven counts and destruction prayer

Theory of harm (allegations). Plaintiffs allege OpenAI and Microsoft copied Seattle Times and Newsday journalism without permission to train and operate generative AI products, including ChatGPT, Copilot, and Bing Chat. The complaint frames those products as "answer engines" whose purpose is to "regurgitate, summarize, and mimic the expressive voice" of the news works they depend on, keeping users on the AI interface instead of visiting seattletimes.com or newsday.com.

Verbatim-output examples (allegations). The complaint alleges model testing reproduced substantial passages of plaintiffs' articles verbatim from prompts that supplied only headline, date, and URL. One pleaded example claims ChatGPT regurgitated 88 consecutive words from a Seattle Times Pulitzer-winning Boeing 737 Max series. Side-by-side Newsday passage tables appear in the PDF as in-body comparisons of alleged regurgitation (separate from the registered-works and trademark exhibits). Those tables are plaintiff illustrations. They are not adjudicated findings.

CMI and RAG (allegations). Plaintiffs plead that defendants removed copyright management information (CMI) such as titles, authors, and copyright notices, and that retrieval-augmented generation continues to make new copies of current content without preserving CMI.

Referral-traffic statistic as pleaded. Citing industry data for December 2024 to December 2025, the complaint alleges search referral traffic to mid-sized regional and metro daily publishers "like Newsday and The Seattle Times" declined by approximately 47%, versus about 22% for larger national publishers. The complaint attributes the steeper midsize decline to dependence on incidental search traffic that AI-generated answers now keep. The figure is pleaded from secondary industry sources named in the footnotes. It is not a court finding about these plaintiffs' own logs.

Seven counts.

  1. Copyright infringement (17 U.S.C. § 501)
  2. Vicarious copyright infringement
  3. DMCA § 1202(b)(1) (removal of CMI)
  4. DMCA § 1202(b)(3) (distribution of works with CMI removed)
  5. Federal trademark dilution (15 U.S.C. § 1125(c))
  6. Washington trademark dilution (RCW 19.77.160)
  7. New York trademark dilution (N.Y. Gen. Bus. L. § 360-L)

Prayer for relief. Plaintiffs seek statutory or actual damages and profits under the Copyright Act and DMCA; permanent injunctions; and, under 17 U.S.C. § 503, "impoundment and/or destruction" of copies of plaintiffs' works "and all LLMs and training datasets incorporating Plaintiffs' works or derivatives thereof" in defendants' possession, custody, or control; plus trademark injunctions, accounting, treble damages as pleaded, fees, costs, and other relief. Dated September 4, 2026.

Microsoft-OpenAI investment facts as pleaded. The complaint recounts Microsoft's investments in OpenAI (including an alleged ~$13 billion commitment and valuation figures drawn from Microsoft and OpenAI public statements). It does not state that OpenAI or Microsoft funded The Seattle Times or Newsday as newsrooms.

Midsize publishers join the S.D.N.Y. wave

This filing adds two regional/metro publishers to the S.D.N.Y. copyright and DMCA path already occupied by larger outlets in the OpenAI MDL. The complaint's distinctive operator-facing elements are the "answer engine" product framing, the pleaded 47% midsize referral decline, the CMI counts, state and federal trademark-dilution counts, and the § 503 destruction ask aimed at datasets and models.

No answer, motion to dismiss, or consolidation order appears on the face of Document 1. Secondary social chatter that OpenAI or Microsoft "funded" The Seattle Times is not a fact stated in this complaint.

Operator checklist from Document 1

  • Treat every factual paragraph as allegation. Preserve the complaint. Do not brief boards as if infringement or CMI removal were proven.
  • Inventory training and RAG provenance. If you ship answer-style products over news corpora, map whether Seattle Times or Newsday URLs appear in crawl, license, or RAG indexes.
  • Separate copyright, DMCA CMI, and trademark tracks. Dilution counts under federal, Washington, and New York law are distinct from § 501. Brief trademark counsel if product UIs surface plaintiff marks in generative answers.
  • Stress-test destruction exposure. The § 503 ask targets LLMs and training datasets incorporating plaintiffs' works. Model ops and legal need a shared inventory of what "incorporating" would mean for your stack if a court entertained that relief.
  • Read the 47% figure as pleaded industry data. It is a midsize-publisher market claim with cited secondary sources. Pull your own referral logs before using it in a board deck.
  • Do / buy / ignore / distrust. Do: add Case 1:26-cv-07644 to the OpenAI publisher-litigation watch list and calendar the answer deadline once served. Buy: a targeted license or exclusion review for Seattle Times and Newsday content in training and RAG. Ignore: unverified social claims that Microsoft or OpenAI bankrolled these plaintiffs' newsrooms. Distrust: any claim that this complaint alone forces dataset destruction tomorrow.

Allegations we are not treating as proof

  • That defendants in fact copied every Seattle Times or Newsday registered work into training sets. That is pleaded on information and belief.
  • That the 88-word Boeing example, or the Newsday side-by-side tables, establish systemic infringement as adjudicated fact.
  • That the 47% midsize referral decline was caused solely by ChatGPT or Copilot, or matches these plaintiffs' internal analytics. The complaint cites industry sources.
  • That CMI removal was intentional as a proven state of mind. Intent is pleaded.
  • That trademark dilution under federal, Washington, or New York law has been shown. Those are counts for adjudication.
  • That OpenAI or Microsoft funded The Seattle Times or Newsday. The complaint discusses Microsoft investment in OpenAI. It does not plead newsroom funding of these plaintiffs. Social chatter to that effect is outside this record.
  • Any defendant answer. Document 1 is the complaint only.

What Document 1 does not include

  • Proof of service, assigned judge initials beyond the case number stamp, or a consolidation order into 25-md-3143.
  • OpenAI or Microsoft responsive pleadings.
  • Plaintiffs' internal referral-traffic and subscription exhibits beyond the industry citations in footnotes.
  • A court ruling on fair use, DMCA CMI, trademark dilution, or § 503 destruction.
  • License-negotiation history between these plaintiffs and these defendants, if any exists outside the complaint's general market allegations.
  • Independent forensic extraction of the PDF exhibit images for republication rights.

Monday action: Publisher-side and AI-product counsel should archive the 4 Sep 2026 complaint, brief the seven counts and dataset/model destruction prayer as alleged exposure, and open a provenance check for Seattle Times and Newsday text in training and RAG. Do not treat the complaint narrative or the 47% figure as adjudicated fact.