CNN filed suit against Perplexity AI on Thursday, May 28, in the U.S. District Court for the Southern District of New York, alleging the answer engine scraped more than 17,000 stories, photos, and videos from the network without a license and then used that material to power its products.
The 54-page complaint carries two distinct legal theories. The first is standard copyright infringement: Perplexity ingested CNN content, the network says, even after negotiations over a licensing deal broke down and CNN blocked Perplexity’s crawler from accessing its site. The second theory is more operationally specific. CNN alleges that Perplexity’s “Comet Plus” subscription tier advertised access to CNN’s premium content despite no commercial relationship existing between the two companies. That claim falls under trademark law, and it goes beyond the scraping question: Perplexity, CNN argues, was actively trading on the network’s brand to sell paid subscriptions.
Variety, which first reported the filing, quoted Perplexity’s full response as four words: “You can’t copyright facts.” That defense is familiar in AI copyright litigation. It draws on the idea-expression dichotomy, which holds that raw facts are not protectable, only the selection and arrangement of them. Courts have not yet resolved how that principle applies to large-scale automated ingestion of journalistic work, and no U.S. appellate decision has settled the question.
CNN is not the first news company to sue Perplexity. The New York Times, Dow Jones (publisher of the Wall Street Journal), and the New York Post have filed comparable suits. What distinguishes the CNN action is the combination of broadcast scale and trademark allegations tied to a specific paid product feature. A subscription tier that names a publisher as a content source, when that publisher has explicitly refused to license its material, creates a narrower and potentially more provable harm than broad copyright claims over training data.
The lawsuit also arrives at a moment when Perplexity is expanding its publisher relationships in two directions at once. Time and USA Today’s parent company (formerly Gannett) have struck licensing deals with the company. CNN separately concluded a content-licensing arrangement with Meta in December 2025. The diverging outcomes illustrate that the industry has not settled on a single model: some publishers are licensing, some are litigating, and the legal framework for AI-driven answer engines is being written case by case in federal courts.
For search and content teams, the trademark portion of this complaint is the detail worth tracking. If courts find that an AI platform cannot promote publisher-branded content tiers without a formal agreement, the liability exposure for answer engines expands well beyond training-data questions into the product and marketing layer. That would require AI search platforms to audit not just what content their models ingested, but how their subscription and feature copy describes source relationships. Publishers who have not yet formalized their relationship with Perplexity, positively or negatively, should treat the Comet Plus allegation as a prompt to document their status clearly.
CNN is seeking statutory damages and a court order barring Perplexity from using its content or marks.
Reported by Variety on 2026-05-28, written by Corbin Bolies.
