Lawsuit tracker
NYT v. OpenAI
The New York Times Co. v. Microsoft Corp. and OpenAI, No. 1:23-cv-11195 (S.D.N.Y.), part of In re: OpenAI, Inc. Copyright Infringement Litigation (MDL No. 3143)
| Plaintiffs | The New York Times Company |
|---|---|
| Defendants | OpenAI entities and Microsoft |
| Court | U.S. District Court, Southern District of New York (Judge Sidney H. Stein; discovery before Magistrate Judge Ona T. Wang) |
| Filed | 2023-12-27 |
| Status | Active |
| Content type | journalism |
| Last updated | 2026-07-21 |
The claims
Copyright infringement (direct and contributory), DMCA copyright-management-information claims, unfair competition and trademark dilution. The Times seeks damages it puts in the billions, plus destruction of models and training sets built on its work.
What has happened
The Times was the first major American newspaper to sue over generative AI. It alleges OpenAI and Microsoft copied millions of its articles to build GPT models, and its complaint included roughly one hundred exhibits of ChatGPT reproducing Times journalism nearly verbatim. OpenAI answered that training is fair use and accused the paper of contrived prompting. Judge Stein let the core copyright claims through in spring 2025, and the case was folded into a multidistrict litigation gathering the federal OpenAI copyright suits — newspapers and authors alike — for pretrial proceedings. Since then the fight has been about discovery: a 20-million-log production order affirmed in January 2026, orders reaching a co-founder’s journal and unprepared corporate witnesses, and a July 2026 sanctions motion accusing OpenAI of concealing and deleting evidence.
Key developments
- 2023-12-27 — The Times sues OpenAI and Microsoft in Manhattan federal court, attaching side-by-side exhibits of ChatGPT outputs that closely track its articles.
- 2025-03-26 — Judge Stein largely denies the motions to dismiss, keeping the core copyright claims — including contributory infringement — in the case while trimming some peripheral theories. A written opinion follows on April 4, 2025.
- 2026-01-05 — Judge Stein affirms Magistrate Judge Wang’s order requiring OpenAI to produce a sample of about 20 million de-identified ChatGPT conversation logs, rejecting OpenAI’s user-privacy objections; the logs bear on outputs and on the fair use defense.
- 2026-04 — After finding OpenAI’s corporate designee unprepared on noticed topics, Magistrate Judge Wang orders OpenAI to produce a properly prepared witness for re-deposition and gives plaintiffs additional deposition time.
- 2026-07-09 — The Times, the Daily News plaintiffs, CIR, The Intercept and Ziff Davis move for sanctions, alleging OpenAI concealed an internal database of roughly 78 million de-identified chats it used to assess infringement and deleted logs covered by preservation orders. The motion is pending.
Why it matters for training data
This is the case setting the discovery playbook for AI litigation: usage logs, training-corpus audits and executives’ notes are all discoverable, and retention practices have become litigation infrastructure. For data buyers and suppliers, whatever fair-use ruling eventually issues here will effectively price news content for every model builder — and the sanctions fight shows courts will police how AI companies handle evidence about their own training data.
Sources
- S.D.N.Y. opinion on motions to dismiss (April 4, 2025)
- ABA Journal: 20M chat logs must be turned over
- TechCrunch: NYT says OpenAI hid evidence (July 2026)
- Variety: outlets seek sanctions against OpenAI
Deeper analysis
Want data that clears this in diligence?
Whether you're building a model or sitting on an archive, the first conversation is short and specific.
Send a brief