images.ctfassets.net · Original source page
Image provenance
Inherited source visual. Image capture date and exact event relationship were not established again in this expansion. Owner publication review pending; credit does not grant permission.
Original assetThe signal
On 27 December 2023 The New York Times, a publishing-sector news organisation, filed a copyright complaint in the US District Court for the Southern District of New York against Microsoft and several OpenAI entities. The docket, case number 1:23-cv-11195, records a copyright-infringement claim: the complaint alleges the defendants used the Times' copyrighted articles without authorisation or payment to train and commercialise AI models, and that those models can reproduce close to verbatim excerpts of Times journalism. These are allegations set out in a filed complaint, not judicial findings.
The evidence
OpenAI responded publicly on 8 January 2024 in a post titled "OpenAI and journalism", calling the lawsuit "without merit" and saying discussions with the Times had been "progressing constructively" before the complaint was filed. The company argues that using "publicly available internet materials is fair use," that Times content "didn't meaningfully contribute" to training its existing models, and that close reproductions of Times text came from users who pasted lengthy article excerpts to prompt the model into reproducing them, rather than from the model surfacing that text unprompted. OpenAI also points to licensing partnerships with the Associated Press and Axel Springer as evidence it will pay for content on negotiated terms. Both documents describe the same dispute from opposite sides: one sets out allegations, the other a public rebuttal; neither is a court ruling.
Timeframe and confidence
The docket record and OpenAI's statement are each dated and consistent on the underlying facts, that a complaint was filed 27 December 2023 alleging copyright infringement over AI training data. This is editorial reading: a filed lawsuit is a genuine signal that a licensing dispute has become formal and public, but the substantive question, whether training on copyrighted text without a licence is infringement or fair use, remains contested and undecided at the point these documents were produced.
What would change the reading
A ruling on the merits, a settlement with disclosed licensing terms, or a broader licensing market between publishers and AI developers, several of which OpenAI's statement points to as already under way, would each resolve or reshape what this filing signals about how AI training data gets priced.
- Did courts treat training on copyrighted text as fair use, infringement, or something requiring case-by-case analysis?
- Did more publishers sign licensing deals with AI developers rather than litigate?
- Did the specific verbatim-reproduction evidence cited in the complaint hold up as the case proceeded?
A single lawsuit does not settle a legal question, but a major publisher choosing litigation over further negotiation is itself evidence that a licensing market for AI training data had not yet formed on terms both sides would accept.
Source trail
- The New York Times Company v. Microsoft Corporation, Docket No. 1:23-cv-11195 (S.D.N.Y.)www.courtlistener.com · Source publication: not established · Retrieved 2026-09-16
Records the case caption, court, filing date and copyright-infringement cause of action underlying the New York Times' complaint against Microsoft and several OpenAI entities.
- OpenAI and journalismopenai.com · Source publication: 2024-01-08 · Retrieved 2026-09-16
Records OpenAI's public response to the lawsuit, calling it without merit and setting out the company's fair-use argument and news-publisher partnerships.
- Event date
- 2023-12-27
- First source date
- 2024-01-08
- Source-record publication
- Not supplied — draft retained
- Preparation
- 2026-09-16