CNN is suing Perplexity AI. The lawsuit, *Cable News Network Inc. v. Perplexity AI, Inc.*, No. 1:26-cv-04427 (S.D.N.Y.), was filed May 28, 2026, and names both copyright and trademark infringement claims. The case is independently verifiable via PACER, the case number is included here so compliance and legal teams can access the filing directly.
The complaint reportedly identifies two Perplexity crawlers by name, “PerplexityBot” and “Perplexity-User”, as the mechanism by which CNN’s content was accessed and reproduced. According to trade press coverage of the filing, CNN alleges its complaint cites more than 17,000 distinct copyrighted works, including articles, videos, and images. All claims about the complaint’s contents are allegations, no court has made any findings.
The licensing allegation is the legally significant detail. CNN states in the complaint that it attempted to negotiate a commercial licensing agreement with Perplexity in late 2025. Perplexity declined. CNN’s legal theory, according to coverage of the filing, is that this refusal establishes the company’s prior knowledge that continued scraping was unauthorized, a predicate for willful infringement, which carries significantly higher statutory damages under 17 U.S.C. § 504(c)(2).
This matters because Perplexity isn’t a small target anymore. The company has accumulated an active multi-plaintiff docket. Earlier AI copyright enforcement actions targeted smaller platforms; CNN brings global brand recognition, substantial legal resources, and a journalism industry with acute economic motivation to pursue this litigation through trial. The complaint also alleges Perplexity’s outputs are “identical or substantially similar” to CNN’s proprietary content, language that directly invokes the reproduction standard in copyright law.
The catch is that CNN is the fourth major plaintiff to name Perplexity specifically. The New York Times, Reddit, and Yomiuri Shimbun have all reportedly filed or threatened similar claims. The pattern across these cases is consistent: publishers allege scraping without licensing, prior contact establishing knowledge, and outputs that substitute for original content. Perplexity’s core product, the “answer engine” that synthesizes web content into direct responses, is the liability mechanism itself.
What to watch
Perplexity hasn’t responded to the complaint as of the filing date. Its answer or motion to dismiss will reveal its primary defense theory. Two paths are possible. It contests the copying (factual defense: the crawlers didn’t take what CNN says they took). Or it contests the legal theory (a fair use argument that synthesis constitutes transformative use). A fair use defense would have far broader implications for the AI industry, a ruling either way on that question shapes every RAG-based product that ingests copyrighted content.
What to Watch
The real question is whether the accumulating docket triggers a licensing negotiation rather than a series of trials. Four named plaintiffs pursuing coordinated litigation against a single defendant creates settlement pressure that individual cases don’t. Coverage of the filing doesn’t indicate any settlement discussions, but the economics of serial litigation typically favor structured resolution over prolonged defense.
Don’t expect this to move quickly. Federal copyright cases in S.D.N.Y. routinely take 18 to 36 months before trial. But compliance teams building AI content pipelines shouldn’t wait for a verdict. The failed licensing allegation is the most transferable signal here: if your product ingests content from publishers you’ve previously contacted about licensing and declined, you’ve potentially established the knowledge element for willful infringement regardless of how the Perplexity cases resolve.