
As of November 10, 2026, the case is in discovery before Judge Failla. The court denied Perplexity's motion to dismiss or transfer in full on August 21, 2025, and no court has ruled on infringement or fair use.
This page is part of the independent reference on AI training data licensing, and DataPacts is not affiliated with any party and gives no legal advice.
All dates and docket entries here come from the court documents, and how figures are sourced and dated is set out on the methodology page.
Dow Jones v. Perplexity at a Glance: Court, Docket, Parties and Status
The case is past the pleading challenge and in discovery, with no ruling on the merits as of November 10, 2026.
The plaintiffs are the publisher of The Wall Street Journal and the publisher of the New York Post, and the defendant is the company behind the Perplexity answer engine. The table lists the case facts as of 2026-11-10 with the court document behind each.
| Field | Value | Source |
|---|---|---|
| Case | Dow Jones & Company, Inc. and the publisher of the New York Post v. Perplexity AI, Inc. | Doc. 1 |
| Court | U.S. District Court, S.D.N.Y. | Doc. 1 |
| Docket | 1:24-cv-07984-KPF (caption of Doc. 1: 24-cv-7984) | Doc. 227 |
| Judge | Katherine Polk Failla, U.S. District Judge | Doc. 227 |
| Filed | 2024-10-21 | Doc. 1 (ECF stamp) |
| Plaintiffs | Dow Jones & Company, Inc. (The Wall Street Journal); the New York Post's publisher (New York Post) | Doc. 1 |
| Defendant | Perplexity AI, Inc. | Doc. 1 |
| Claims | 17 U.S.C. 106 (two counts); 15 U.S.C. 1125 (one count) | Doc. 1 |
| Key ruling | 2025-08-21: motion to dismiss or transfer denied in full | Doc. 65 |
| Status | discovery; motion to compel granted in full 2026-10-29 | Doc. 227 |
| Settlement | none recorded | case record |
What Dow Jones and the New York Post Claim Against Perplexity
The publishers plead three counts: copyright infringement through the copies Perplexity makes for its retrieval index, copyright infringement through answers that reproduce or summarize their articles, and false designation of origin and trademark dilution when answers attribute invented text to their mastheads.
The complaint gives examples, all allegations: a Perplexity Pro user receiving the full text of a New York Post column, a free-tier summary of a paywalled Wall Street Journal article, and an answer that reproduced 139 words of a Post article and added five invented paragraphs.
| Count | Legal basis | What the publishers say Perplexity copied or did | Remedy sought | Source |
|---|---|---|---|---|
| One | Copyright infringement, 17 U.S.C. 106 | copies of articles as "inputs" for its retrieval index, from their sites, licensees' sites and third-party indices | injunction; destruction of any index holding the works; statutory or actual damages and profits | Doc. 1, pp. 34 to 35 |
| Two | Copyright infringement, 17 U.S.C. 106 | "outputs": answers with verbatim or near-verbatim text, summaries or abridgements | same, per output copy | Doc. 1, pp. 35 to 37 |
| Three | False designation of origin and dilution, 15 U.S.C. 1125(a)(1) and (c) | answers attributing invented text to The Wall Street Journal, WSJ and New York Post marks | injunction; damages up to three times actual damages, profits | Doc. 1, pp. 37 to 39 |
Allegations from the complaint of 2024-10-21; no count has been decided as of 2026-11-10.
The prayer for relief in the complaint of October 21, 2024 asks for injunctions, destruction of any index holding the works in any format, and statutory damages up to USD 150,000 per infringement, which is the willful ceiling in 17 U.S.C. 504(c). It also demands a jury trial.
The same prayer ties the Lanham Act count to damages of up to three times actual damages, so the three counts carry different remedies: statutory damages per infringed work for the copyright counts and enhanced actual damages and profits for the trademark count.
None of these remedies has been awarded, and the court has not yet ruled on any count.
Why the Complaint Targets Retrieval Copies, Not Model Training
The complaint pleads that assembling sources for a retrieval-augmented generation (RAG) index is a process distinct from assembling content to train or fine-tune a model, and its counts rest on the index and the answers, not on training.
That is why the parent tracker lists the case among related proceedings and not among the training cases.
The publishers allege that Perplexity crawls their sites with its own crawlers, sometimes ignoring or evading robots.txt, and also copies from licensees' sites and third-party crawlers.
Perplexity's own documentation, checked 2026-11-09, says PerplexityBot "is not used to crawl content for AI foundation models" and that Perplexity-User "generally ignores robots.txt rules"; the crawler statements and the stealth-crawling dispute belong to the page on PerplexityBot and Perplexity-User.
Cloudflare reported on 2025-08-04 that an undeclared crawler made 3 to 6 million requests per day and removed Perplexity from its verified bot list. That is Cloudflare's own measurement, offered here as context, and it is not a finding in this case.
The distinction matters for licensing: data acquired for retrieval and answers sits next to training data, and display or answer licenses differ from training licenses.
The parties' public positions and pleadings raise fair use, but it is undecided here, and fair use and AI training in US law sets out the four factors and the rulings so far.
Dow Jones v. Perplexity Docket Entries, 2024 to 2026
Nine docket entries carry the case from complaint to the current discovery dispute: the complaint, the second amended complaint, Perplexity's motion to dismiss or transfer, the protective order, the opinion denying that motion, the scheduling order, the order compelling discovery and two non-party letters seeking the sealed order, both granted.
The table lists them as of 2026-11-10, with the source of each. Two entries, the Second Amended Complaint and the motion, are known from the later opinion and protective order that refer to them, so their filing dates are not stated here. The other seven carry the date shown on the court document.
| Date | Doc. | Entry | What it decided or reported | Source |
|---|---|---|---|---|
| 2024-10-21 | 1 | Complaint | three counts; jury demanded | Doc. 1 |
| not recorded here | 46 | Second Amended Complaint | the pleading Perplexity moved against | Doc. 65 |
| not recorded here | 47 | Perplexity's motion to dismiss or, in the alternative, to transfer venue | pending on 2025-08-01 | Doc. 62, n.1 |
| 2025-08-01 | 62 | Confidentiality and Protective Order, so ordered | "Inspection Data" tier for source code and training data | Doc. 62 |
| 2025-08-21 | 65 | Opinion and Order | motion to dismiss or transfer denied in full; ten later works kept | Doc. 65 |
| 2025-08-29 | 68 | Civil Case Management Plan and Scheduling Order | schedule dates not recorded here | Doc. 68 |
| 2026-10-29 | 227 | Order | motion to compel (Dkt. 218) granted in full; reasons in sealed order Dkt. 226 | Doc. 227 |
| 2026-10-30 | 229 | Non-party letter (Chicago Tribune, The New York Times, Wirecutter, The Athletic), so ordered | access to the sealed order, unopposed by both parties | Doc. 229 |
| 2026-11-06 | 231 | Non-party letter (Britannica, Merriam-Webster), memo endorsed | request to receive the sealed order Dkt. 226 granted, unopposed by both parties | Doc. 231 |
The August 21, 2025 Opinion: Perplexity Stays in New York
Judge Katherine Polk Failla's Opinion and Order of August 21, 2025 (Doc. 65) denied in full Perplexity's motion to dismiss the Second Amended Complaint for lack of personal jurisdiction and improper venue or, in the alternative, to transfer the case to the Northern District of California.
The motion also sought dismissal under Rule 12(b)(6) as to ten works, and the court denied that part as well.
The ten works were registered after the initial complaint, which covered 326 registered works, and the first amended complaint covered 336. The court held the ten timely as to the amended pleading.
The opinion asked the parties for a case management plan, and the scheduling order followed on August 29, 2025 (Doc. 68). It decides where and whether the case proceeds, not whether Perplexity infringed.
In practice the ruling means that the publishers' choice of New York stands: Perplexity argued that the case belonged in California, and the court kept it in the district where the plaintiffs sued.
The merits questions, including whether copying for a retrieval index or an answer is infringement or fair use, were not before the court on that motion.
Discovery in 2026 and the Sealed Order of October 29
On October 29, 2026, the court granted the publishers' motion to compel discovery in full (Doc. 227), giving its reasons in a sealed order (Dkt. 226) that resolved the motion at Dkt. 218. What the sealed order says is not public.
The protective order of August 1, 2025 (Doc. 62) sets an "Inspection Data" tier for source code and other sensitive electronic data "including but not limited to training data", inspected under a separate agreement. Perplexity reserved its venue objection in a footnote while its motion was pending.
Non-parties in other Perplexity suits then asked for the sealed order. Counsel for the Chicago Tribune, The New York Times, Wirecutter and The Athletic filed a letter on October 30, 2026 that the court so ordered (Doc. 229), and the court granted the same access to Britannica and Merriam-Webster on November 6, 2026 (Doc. 231).
No trial date is recorded in the sources read for this page, and the sources do not show that fact discovery has closed.
What the Case Means for Holders of News Content Weighing a License
The complaint frames the dispute as a refused license: the publishers say they wrote to Perplexity in July 2024 to raise the legal issues and offer to discuss a licensing deal, received no response, and point to News Corp's partnership with OpenAI as the licensed alternative.
Terms for news publishers licensing content to AI companies are covered on the publishers page. News Corp's OpenAI partnership of 2024-05-22 lets OpenAI display content in answers, and the parties did not disclose its value.
Perplexity's own licensed routes in the record each have a stating party and a date. Le Monde announced an agreement on 2025-05-14 for answers only, and in Le Monde's words it does not provide for using its content to train a model.
Gannett's chief executive described a Perplexity agreement as "fair value of our content" on 2025-07-31 without an amount, and Getty Images announced a display license on 2025-10-31. Perplexity's publisher revenue-share terms are not verified, so no percentage or pool size is given.
For a holder, the trade-off is plain. A license for answers can bring revenue and attribution terms, while a suit brings cost and years of discovery, as this docket shows. The deal tracker of AI data licensing deals dates each Perplexity deal with the party that stated it.
Other Lawsuits Against Perplexity
Dow Jones is the earliest of the suits against Perplexity recorded on the case tracker, grouped here in five rows, and all of them concern copying for AI search answers or circumventing access controls rather than training a model.
The parent case tracker of AI copyright lawsuits over training data lists them as related proceedings outside the training tracker, and the table gives each with its status as of 2026-11-09 (this case as of 2026-11-10).
| Case | Court and docket | Claim type | Status | Page |
|---|---|---|---|---|
| Dow Jones and NY Post v. Perplexity | S.D.N.Y. 1:24-cv-07984-KPF | copyright and trademark, answer engine | discovery; motion to compel granted 2026-10-29 | this page |
| Britannica and Merriam-Webster v. Perplexity | S.D.N.Y. 1:25-cv-07546-JLR-SLC | copyright and trademark, answer engine | discovery; 2026-10-27 order (Doc. 152) on "Material Excluded" | linked case page |
| News publishers v. Perplexity (NYT, Chicago Tribune, CNN) | S.D.N.Y. 1:25-cv-10106-LAP, 1:25-cv-10094-LAP, 1:26-cv-04427-LAP | copyright (NYT also trademark) | pending before Judge Preska; NYT and Tribune stipulated source code protocol submitted 2026-10-21; CNN motion to dismiss fully briefed, oral argument requested 2026-10-23 | linked case page |
| Reddit v. Perplexity, SerpApi and others | S.D.N.Y. 1:25-cv-08736-PAE (one case) | DMCA 1201 circumvention over search-result scraping | motions to dismiss of SerpApi and Perplexity predominantly denied 2026-07-31 (Dkt. 104) | linked case page |
| Yomiuri v. Perplexity | Tokyo District Court | AI search answers | not stated | linked case page |
Perplexity's crawler statements, licensing deals and program terms are covered on Perplexity as a data buyer. Suits over Perplexity's shopping agent or its users' privacy are outside this reference. Each row states the court and docket from the filings, and the table is refreshed when a docket entry changes a status.
More About the Dow Jones Lawsuit
Allegations Versus Rulings: What Is Decided and What Is Still Open
Only procedure is decided: the court has kept the case in New York with all 336 works and has compelled discovery, while every infringement, fair use and trademark question is still open.
Decided
- Jurisdiction and venue (Doc. 65)
- No transfer to the Northern District of California (Doc. 65)
- Ten later-registered works stay in the case (Doc. 65)
- Motion to compel discovery granted (Doc. 227)
Still open
- Infringement by index copies
- Infringement by answers
- Fair use
- Trademark claims
- Remedies
- Trial date
Frequently Asked Questions About the Dow Jones Lawsuit
Is Dow Jones v. Perplexity about AI training?
No, the counts target copies in a retrieval index and in answers. The complaint itself separates assembling a RAG index from training a model.
Has the court ruled on fair use?
No, the court has not ruled on fair use as of 2026-11-10. The only rulings concern jurisdiction, venue, transfer, the ten later works and discovery.
Was the case moved to California?
No, the case stays in New York. On 2025-08-21 the court denied transfer to the Northern District of California (Doc. 65).
Is it legal to scrape news sites for AI answers?
This case has not decided that question. Neither the complaint nor any ruling so far resolves whether collecting news content for AI answers is lawful.
The copyright, contract and computer-misuse rules are set out on the page on web scraping for AI and the law.
Has Perplexity settled with Dow Jones?
No settlement is recorded as of 2026-11-10. The case is in discovery and no trial date is recorded in the sources read.