New

NYT Sues Perplexity AI Over Alleged Copyright and Trademark Violations

The New York Times filed suit in the Southern District of New York alleging that Perplexity AI copied, distributed, and displayed millions of Times articles without permission to fuel its generative AI products. The complaint also points to fabricated outputs being presented alongside NYT trademarks, creating the impression the newspaper authored or endorsed the material. The Times seeks damages and injunctive relief.

The case lands amid a wave of litigation against the San Francisco startup, which is reportedly valued near $20 billion. Recent plaintiffs include the Chicago Tribune, Reddit, Encyclopedia Britannica, Dow Jones, and the New York Post. The Times previously sent a cease-and-desist notice more than a year ago.

What the Times Alleges

• Systematic scraping and copying of content, including paywalled material.
• Use of NYT trademarks alongside AI-generated "hallucinations," implying source or endorsement.
• Commercial use of Times content to build and promote Perplexity's products.

The newspaper's position is straightforward: unlicensed use for product development and marketing violates its rights and causes market harm.

Perplexity's Response

Perplexity disputes the claims, characterizing the lawsuits as a tactic used against new technologies. The company says it indexes web pages and provides citations, and that it is not scraping data for building foundation models. Expect arguments that outputs are transformative and that any ingestion falls within existing norms of web indexing.

Broader Context

This dispute is part of an ongoing fight over the use of copyrighted text to train or power AI systems. Last year, it was reported that multiple AI companies bypassed a web standard intended to block scraping. That standard-robots.txt-reflects preference, not legal consent, though it can be evidence in contract and notice disputes. For reference, see the Robots Exclusion Protocol specification at RFC 9309.

The Times has licensed some content to technology companies (e.g., for voice assistants) and is also in conflict with other AI providers. NYT shares ticked up about 1.8% after the news.

Key Legal Questions

• Copyright: Does ingestion and output constitute infringement of reproduction, distribution, or public display rights? How do courts weigh market substitution when outputs summarize or restate paywalled content?
• Trademark/Lanham Act: Do hallucinated outputs paired with NYT marks create false designation or endorsement? See 15 U.S.C. § 1125 at LII.
• Contract/TOS: If Perplexity accessed pages in violation of NYT terms, does that support breach of contract or state-law unfair competition claims?
• Anti-circumvention: If paywalls were bypassed, does that trigger DMCA §1201 liability (and related claims over removal or alteration of copyright management information)?
• CFAA and scraping: hiQ v. LinkedIn narrowed CFAA risks for publicly accessible pages, but paywalls and authenticated areas remain a live wire. SDNY will apply Second Circuit precedent, so expect briefing on circuit splits and the public/private access line.
• Fair use: Perplexity may argue transformative use, functional indexing, or quotation for reference. The Times will emphasize the commercial nature, the qualitative value of the excerpts, and substitution effects.

Injunctive Relief Posture

NYT is seeking to stop alleged use while the case proceeds. The court will evaluate likelihood of success, irreparable harm, balance of equities, and public interest. Evidence showing direct substitution of paywalled content or source confusion tied to trademarks could weigh heavily on the first two factors. Preservation orders for logs, training data, and model versions are likely early asks.

What In-House and Outside Counsel Should Do Now

• Inventory: Map how your AI stack ingests, stores, and uses third-party content (including paywalled and user-generated material).
• Consent and controls: Align crawling and ingestion with explicit licenses; enforce robots and publisher signals consistently; gate access to paywalled content.
• Attribution and brand safety: Prevent outputs from pairing third-party marks with unverifiable claims; tighten disclaimers and provenance labeling.
• Recordkeeping: Retain crawl logs, dataset lineage, model versioning, and output audits. You'll need this for injunction defense or negotiation.
• Terms of use hygiene: Ensure compliance with site terms and API agreements; route all exceptions through legal approvals.
• Risk triage: Prioritize high-value publishers and sensitive verticals (news, finance, health). Consider targeted licenses where substitution risk is high.

What to Watch Next

• Early motions on preliminary injunction and expedited discovery.
• How the court frames fair use for AI systems that both index and summarize paywalled content.
• Whether trademark theories tied to hallucinations gain traction and spur broader brand-usage constraints.
• Any coordination among publisher plaintiffs or consolidation attempts.

For legal teams building internal AI literacy and governance programs, a structured training path helps you spot these risks earlier. See curated AI learning tracks by role at Complete AI Training.