Perplexity AI Says 'You Can't Copyright Facts' in Defense Against CNN Copyright Suit

A single legal argument — that facts cannot be copyrighted — could determine the fate of hundreds of AI copyright cases. Perplexity AI, a rising search engine powered by large language models, is telling a federal court that CNN's copyright suit against it should be thrown out because the network cannot own the factual information its news articles contain. The case, filed in 2025, is one of more than 100 copyright lawsuits brought against AI companies as of early 2026, and its outcome could reshape how generative AI models are trained and deployed.

Perplexity AI filed a motion to dismiss CNN's complaint in the U.S. District Court for the Southern District of New York. The company's core defense hinges on the 'idea–expression dichotomy,' a bedrock principle of U.S. copyright law that protects the specific expression of an idea but not the underlying facts or concepts. Perplexity argues that its AI search engine — which can summarize news articles and answer user queries by pulling information from multiple sources — uses only the factual content of CNN stories, not their creative expression.

The lawsuit is part of a wave of litigation that has deeply unsettled the artificial intelligence industry. News organizations, authors, visual artists, and music publishers have all sued AI companies over the use of copyrighted material in training datasets or in generated outputs. CNN specifically alleges that Perplexity reproduces substantial portions of its articles without permission, diminishing the value of its journalism and damaging its business model. Perplexity counters that its summaries are transformative, not derivative, and that any copying is fair use.

The Perplexity AI copyright defense has drawn attention because it directly tests the limits of copyright in the age of large language models. If the court accepts that facts are not copyrightable and that AI summarization is a lawful fair use, it could provide a safe harbor for many AI companies. Legal experts point out that the 'idea–expression dichotomy' has long been used to protect news reporting itself — journalists cannot copyright facts like 'the president signed a bill on Tuesday' — but applying that principle to AI-generated summaries is novel.

More than 100 copyright lawsuits have been filed against AI companies as of early 2026, according to data from the Copyright Office. Notable cases include The New York Times's suit against OpenAI, Getty Images' case against Stability AI, and a class action by authors including John Grisham and George R.R. Martin. The Perplexity AI defense, if successful, could influence how those cases are resolved, though each involves different factual contexts and types of copyrighted works.

Broader implications extend beyond the courtroom. If courts rule that AI systems can freely use facts from copyrighted sources, news publishers may lose a key revenue stream — licensing fees from AI companies. Several major publishers, including News Corp, Axel Springer, and The Associated Press, have already struck licensing deals with OpenAI. A ruling that weakens copyright protection for news content could shift the balance of power back toward AI companies while accelerating the adoption of AI search products that compete directly with traditional media.

What happens next? The court will hear arguments on Perplexity's motion to dismiss in the coming months. A decision is expected by late 2026. If the motion is denied, discovery will proceed, potentially revealing how Perplexity trains and operates its models. If granted, it could set a major precedent. Meanwhile, Congress is considering legislation that would require AI companies to disclose their training data and compensate copyright holders, adding another layer of uncertainty. The Perplexity AI copyright defense is more than a legal maneuver — it is a test of whether the foundational principle that 'you can't copyright facts' survives the AI era.