What happened and why it matters
On December 5, 2025, The New York Times (NYT) filed a federal lawsuit in the Southern District of New York alleging that Perplexity AI copied, distributed, and displayed Times journalism “verbatim or near-verbatim” to power its AI search and browsing tools. The complaint also says Perplexity’s products hallucinated content and displayed it alongside NYT trademarks, potentially confusing readers. The Times seeks damages and an injunction stopping further use of its work. Reuters, The Verge, Justia docket.
The core claims vs. Perplexity’s response
Key allegations and responses
| NYT’s allegation | What it means in practice | Perplexity’s response |
|---|---|---|
| Copyright infringement via “verbatim or near-verbatim” reproduction of articles, including paywalled content | The paper argues Perplexity’s RAG-style systems fetch and re-publish protected text without a license | Perplexity says it indexes public webpages to cite factual content, doesn’t train foundation models on publishers’ data, and that its service is akin to a search tool with citations. Reuters |
| Bypassing website controls and ignoring robots.txt | The complaint echoes 2024–2025 reporting that Perplexity accessed content despite “do not crawl” directives | Perplexity has disputed these characterizations; after Cloudflare warned about “stealth” crawling, a spokesperson called the post a “sales pitch,” and the company said agents sometimes fetch URLs on users’ behalf. Ars Technica, WIRED, TechCrunch |
| False association/brand confusion under trademark law | NYT says hallucinated outputs were shown alongside its marks | The company disputes broad wrongdoing and frames publisher suits as resistance to new technology. The Guardian, TechCrunch |
Perplexity’s head of communications characterized publisher litigation as a familiar, ineffective response to new media technologies, while the Times maintains that unlicensed use undermines its subscription and licensing businesses. TechCrunch.
This isn’t the first clash—why the timing is significant
- The Times sent Perplexity a cease-and-desist letter in October 2024 over use of its content in summaries and outputs. CNBC.
- The lawsuit arrives as NYT pursues a separate, high-stakes case against OpenAI/Microsoft; in November, a magistrate judge ordered OpenAI to produce a sample of 20 million de‑identified ChatGPT logs—an order OpenAI is fighting on privacy grounds. Reuters, OpenAI blog.
- In parallel, the Times struck a licensing deal with Amazon in May 2025 to make NYT content available in Alexa and to train Amazon’s proprietary models—evidence of a “license-or-litigate” strategy. Bloomberg, TechCrunch, CNBC.
The widening front: other publishers and platforms
The Times’ suit follows fresh actions from major publishers:
- Chicago Tribune sued Perplexity on December 4, alleging verbatim reproduction and paywall bypass via Perplexity’s Comet browser. TechCrunch.
- Britannica and Merriam-Webster sued Perplexity in September, alleging widespread copying and trademark misuse; Perplexity moved to dismiss. Britannica press release, Justia docket.
- Dow Jones and the New York Post (News Corp.) sued Perplexity in October 2024. CNBC.
- Reddit sued Perplexity in October 2025, alleging “industrial-scale” scraping, including via Google SERPs. AP/ABC News, Ars Technica.
Industry trackers count well over forty ongoing copyright disputes involving AI companies, with media publishers increasingly opting for either licensing deals or litigation. CMS tracker, status roundup.
What Perplexity has changed—and what it hasn’t
In mid‑2024, after reporting from Forbes and WIRED described near‑verbatim reproductions and robots.txt concerns, Perplexity launched a Publishers’ Program to share ad revenue when partner content is cited. In 2025, it added Comet Plus, a $5/month plan that allocates 80% of subscription revenue to participating publishers based on visits, citations, and agent actions. It also announced a multi‑year licensing deal with Getty Images. The Verge, Digiday, Axios, TechCrunch.
Still, Cloudflare publicly accused Perplexity in August of using “stealth” crawlers that evade robots.txt and blocking rules, and WIRED reported AWS scrutinized Perplexity’s practices; Perplexity disputed wrongdoing and said agents may fetch URLs on a user’s behalf, similar to a person opening a link. Ars Technica, WIRED.
The legal questions the court will weigh
- Training vs. retrieval/display: Unlike model‑training claims, NYT’s allegations largely target live retrieval and output (RAG). Courts have long frowned on products that substitute for the original publisher—see Associated Press v. Meltwater (S.D.N.Y. 2013), which rejected fair‑use defenses for a paid clipping service that reproduced news ledes and excerpts. AP release, SDNY blog.
- Robots.txt and access controls: The Ninth Circuit’s hiQ v. LinkedIn line of cases suggests scraping public webpages isn’t a CFAA violation, but contract, trespass, and copyright claims can still bite—especially if paywalls or terms are breached. FindLaw, Loeb & Loeb.
- Trademark/false designation: If a system hallucinates and presents content under a publisher’s brand, Lanham Act theories may come into play. The Guardian.
For AI and automation teams: practical takeaways
If you build AI assistants, browsers, or search experiences, the NYT–Perplexity dispute is a timely checklist of risk controls:
- Respect publisher preferences by default.
- Honor paywalls and robots.txt for automated agents, not just “official” crawlers. Provide an admin switch for stricter compliance modes.
- Separate “open a URL for me” from “crawl and cache it.”
- Treat user‑initiated fetches transparently and avoid silent background crawling that looks like bot behavior.
- Minimize verbatim reproduction; add friction to full‑text answers.
- Favor brief snippets with prominent links; throttle display length for paywalled sources.
- Invest in rights—licenses, revenue shares, and opt‑in pipelines.
- Structured deals (e.g., Getty Images, publisher programs) reduce uncertainty and strengthen the product’s supply chain.
- Keep an audit trail of retrieval and attribution.
- Log what was fetched, when, under which user action, and why a snippet was displayed. This helps resolve disputes and improve ranking.
- Label clearly and avoid brand confusion.
- Distinguish citations from endorsements. Show source domains and logos only in ways publishers have approved.
- Build a quality bar for sources.
- Detect and downrank AI‑generated spam to avoid “second‑hand hallucinations.” Forbes.
What to watch next
- Early motions and any request for a preliminary injunction in NYT v. Perplexity (SDNY No. 1:25‑cv‑10106). Justia.
- Whether courts converge on a “news aggregator” standard for AI outputs—AP v. Meltwater is likely to be cited heavily. AP.
- How Perplexity’s publisher initiatives (ad‑share program and Comet Plus 80% revenue share) evolve—and whether more outlets sign on amid litigation. Digiday, Axios.
- The broader economic context: Perplexity’s valuation (recent reports put it near $20B) and its push to differentiate via the Comet browser. Reuters, TechCrunch.
Sources
- Reuters: NYT sues Perplexity for “illegal” copying (Dec 5, 2025). Link
- The Verge: NYT accuses Perplexity of reproducing articles (Dec 2025). Link
- The Guardian: NYT lawsuit overview and claims (Dec 2025). Link
- Justia: NYT v. Perplexity docket (SDNY 1:25‑cv‑10106). Link
- TechCrunch: Chicago Tribune sues Perplexity (Dec 4, 2025). Link
- Britannica/Merriam‑Webster v. Perplexity announcement (Sep 2025). Link
- Ars Technica on Cloudflare’s “stealth crawler” claims (Aug 2025). Link
- WIRED on AWS scrutiny of Perplexity crawling (Jun 2024). Link
- CMS AI + Copyright case tracker. Link
- AP v. Meltwater background (2013). Link
- NYT–Amazon AI licensing deal: Bloomberg / TechCrunch / CNBC. Bloomberg • TechCrunch • CNBC
- Perplexity publisher programs and Comet Plus: The Verge / Digiday / Axios. The Verge • Digiday • Axios
