NYT Lawsuit: OpenAI Copyright Case, Fair Use, and ChatGPT Logs

The New York Times’ lawsuit against OpenAI, filed in December 2023 and joined by Microsoft as a co-defendant, is the marquee copyright case testing whether training large language models on news articles without permission is lawful. As of mid-2026, the case has survived motions to dismiss on its core claims, been consolidated with eleven other suits into a multidistrict litigation before Judge Sidney H. Stein in the Southern District of New York, and completed summary judgment briefing. A ruling is expected in the third quarter of 2026, with a trial possible in late 2026 or 2027 if the claims survive.¹

What the Times Is Alleging

The complaint, filed December 27, 2023 in the Southern District of New York (Case No. 1:23-cv-11195), accuses OpenAI and Microsoft of building their AI products on what the Times calls a “mass of Times copyrighted content” scraped and ingested without a license.²³

Three harms sit at the heart of the case. The models sometimes “memorize” training data and reproduce Times articles nearly verbatim, letting users bypass the paywall. The complaint attached screenshots of ChatGPT outputting consecutive paragraphs from Times articles almost word for word.⁴ ChatGPT has also attributed fabricated information to the paper, which the Times says damages its credibility. And if AI systems can generate free summaries or reproductions of Times journalism, the paper argues, subscription incentives collapse.³

The Times is seeking “billions of dollars” in statutory and actual damages, a permanent injunction, and destruction of any GPT models and training datasets that incorporate its works.⁵⁶ Under federal law, statutory damages for willful copyright infringement can reach $150,000 per work.¹

OpenAI’s Fair Use Defense

OpenAI’s central argument is that training an AI on copyrighted text is “transformative” fair use. The system learns patterns rather than storing copies, the company says, and produces new outputs instead of substitutes for the originals.⁷ OpenAI points to the Google Books litigation, where courts held that scanning millions of books to build a searchable database was transformative because it served a different purpose than the originals.⁸

On the verbatim reproductions in the complaint, OpenAI has called them a technical “bug” it intends to fix and has argued that the specific examples were produced by users who deliberately manipulated prompts in ways that violate the company’s terms of service. The company also frames training on publicly accessible articles as permissible and characterizes its underlying purposes as research-oriented.⁷

The defense faces headwinds. The Supreme Court’s 2023 decision in Goldsmith v. Warhol narrowed transformative use, holding that the Warhol Foundation’s commercial use of a photograph competed with the original and did not qualify as fair use.⁸ No court has yet ruled on fair use in the Times case itself.⁷

Consolidation Into a Multidistrict Litigation

In April 2025, the Judicial Panel on Multidistrict Litigation folded the Times case together with eleven other copyright suits against OpenAI into a single proceeding, In re: OpenAI, Inc. Copyright Infringement Litigation (MDL No. 25-md-3143), also before Judge Stein. The consolidated docket includes class actions from book authors, suits from news organizations such as the New York Daily News and the Center for Investigative Reporting, and other DMCA-focused actions.⁹ Magistrate Judge Ona T. Wang manages scheduling, discovery, and non-dispositive motions.¹⁰

What Survived the Motions to Dismiss

On April 4, 2025, Judge Stein issued a sweeping ruling on the defendants’ motions to dismiss across the associated cases. The core copyright claims went forward:

Direct copyright infringement claims survived, including allegations tied to conduct more than three years before filing. Judge Stein held OpenAI failed to show the Times knew or should have known about the infringement earlier, and rejected the argument that a 2020 Times article about GPT-3 gave sufficient notice.¹¹¹²
Contributory infringement claims survived as plausibly alleged.¹¹
DMCA claims under Section 1202(b)(1) were partially dismissed, surviving against OpenAI in the Daily News and CIR actions but dismissed without prejudice against Microsoft across the board and against OpenAI in the Times action. All Section 1202(b)(3) claims were dismissed without prejudice.¹¹
Common-law unfair competition and misappropriation claims were dismissed with prejudice.¹¹

A separate October 2025 ruling within the broader MDL found it “sufficient to defeat OpenAI’s motion to dismiss that plaintiffs have alleged some outputs that a reasonable jury could find are substantially similar to plaintiffs’ works.”⁹

The Fight Over ChatGPT Logs

Discovery has been dominated by battles over user conversation data. In May 2025, Magistrate Judge Wang issued a preservation order requiring OpenAI to retain ChatGPT output logs after the Times alleged the company was systematically destroying evidence of infringement. OpenAI objected that the order was disproportionately burdensome, given that the data covered roughly 60 billion conversations. Judge Stein denied the objection on June 26, 2025. The preserved logs cover ChatGPT Free, Plus, Pro, and Team subscribers, as well as API users without “Zero Data Retention” agreements; enterprise and education-tier customers are excluded.¹³

Production of those logs came next. In November 2025, Judge Wang ordered OpenAI to produce a de-identified sample of 20 million ChatGPT conversations, finding them relevant to OpenAI’s fair use defense. OpenAI sought reconsideration, then appealed to Judge Stein, who affirmed the order on January 5, 2026 and rejected the argument that user privacy concerns required a narrower keyword-based approach.¹⁴ A March 2026 order expanded production further, compelling OpenAI to turn over additional reservoirs totaling 78 million and 10 million logs on top of the original 20 million.⁹

Other discovery disputes followed. In February 2026, a magistrate judge ordered OpenAI to respond to inquiries about “Project Giraffe,” which plaintiffs described as an internal effort to identify and block infringing outputs. In March 2026, the court ordered production of an executive’s personal journal along with additional chat logs. In April 2026, Judge Wang found an OpenAI corporate designee unprepared for a deposition on noticed topics and granted plaintiffs additional time to depose a properly prepared witness.¹⁵¹ OpenAI has publicly floated the concept of an “AI Privilege” to shield user chat logs from disclosure, but no court has recognized any such privilege.¹³

Where the Case Stands Now

Summary judgment briefing wrapped up on April 2, 2026. A ruling is anticipated in the third quarter of 2026.¹ In June 2026, OpenAI filed a motion to dismiss certain contributory infringement claims based on the Supreme Court’s decision in Cox v. Sony.¹⁵ No trial date has been set. If the claims survive summary judgment, trial is projected for late 2026 or 2027.

How Parallel AI Copyright Rulings Could Matter

Two decisions from summer 2025 will shape how the fair use fight plays out. In June 2025, Senior U.S. District Judge William Alsup held that Anthropic’s use of copyrighted books to train its AI was “exceedingly transformative” and constituted fair use, but only where copies were legally obtained. He allowed claims involving books downloaded from pirate sites to proceed. The same week, a separate federal judge ruled in Meta’s favor but stressed that plaintiffs need specific evidence of market harm to prevail against AI developers.¹⁶

A proposed $1.5 billion class-action settlement in Bartz v. Anthropic, covering nearly 500,000 books Anthropic allegedly downloaded from pirate sites, was preliminarily approved in late September 2025. A final fairness hearing was scheduled for May 2026, and payouts have not yet been distributed. The settlement requires Anthropic to destroy the pirated copies but grants no license for future use.¹⁷ Where OpenAI got its training data, and whether that data harms the market for Times journalism, are the questions those parallel rulings put squarely at the center of the case now heading toward its summary judgment moment.

1
AI Lawsuit Tracker. New York Times v. OpenAI
2
CourtListener. The New York Times Company v. Microsoft Corporation
3
Harvard Law Review. NYT v. OpenAI: The Times’s About-Face
4
UNC Journal of Law and Technology. Copyright Clash: The New York Times Lawsuit Against Microsoft and OpenAI
5
The New York Times. The New York Times Sues OpenAI and Microsoft Over Use of Copyrighted Work
6
BBC News. New York Times Sues OpenAI and Microsoft Over AI Use of Articles
7
Columbia Law Review. NYT v. OpenAI and Microsoft
8
Wolters Kluwer. Is Generative AI Fair Use of Copyright Works: NYT v. OpenAI
9
Norton Rose Fulbright. AI in Litigation Series: An Update on AI Copyright Cases
10
CourtListener. In Re: OpenAI, Inc. Copyright Infringement Litigation
11
Justia. The New York Times Company v. Microsoft Corporation
12
U.S. District Court, S.D.N.Y. Opinion and Order, NYT v. Microsoft Corp.
13
Nelson Mullins. From Copyright Case to AI Data Crisis: How NYT v. OpenAI Reshapes Data Governance
14
U.S. District Court, S.D.N.Y. Order Affirming Discovery Ruling in NYT v. OpenAI
15
Law360. In Re: OpenAI, Inc. Copyright Infringement Litigation
16
NPR. Federal Rules in AI Company’s Favor in Landmark Copyright Infringement Lawsuit
17
Copyright Alliance. Participating in the Bartz v. Anthropic Settlement