CNN Sues AI Search Company Perplexity Over Stolen News Articles

CNN filed a lawsuit against Perplexity AI on Thursday, claiming the company's AI search tool scraped more than 17,000 of its news articles and copies CNN's writing without permission. This is the latest fight between traditional news organizations and AI companies over who owns the right to use news content.

CNN is not alone. The New York Times filed a similar lawsuit against Perplexity in December 2025.

What CNN Says Perplexity Did

CNN claims that Perplexity's AI produces exact copies of CNN's reporting and shows readers content that should only be available to people who pay for CNN's subscriptions. Perplexity is an AI search engine — think of it as a search tool powered by AI that tries to answer your questions directly, rather than just pointing you to websites.

The company also says Perplexity bypasses CNN's paywall. A paywall is when a news site makes you pay to read articles. CNN needs subscription money to stay in business, and Perplexity's AI is allegedly giving away CNN's work for free.

A Failed Deal Led to the Lawsuit

CNN and Perplexity had actually talked about working together. In October 2025, they discussed a deal where CNN's articles would be available inside Perplexity's premium service, and CNN would get paid. But the talks fell apart in November 2025, and now they are fighting in court instead.

This history matters. It shows CNN was willing to let Perplexity use its content — as long as CNN got paid. That could make it harder for Perplexity to argue in court that using CNN's work without permission was okay.

This Is Not a New Fight

We have seen similar battles before. When Google first started scanning the entire web in the early 2000s to build its search engine, newspaper publishers were upset too. They worried Google was stealing their content. Eventually, Google and news companies worked out deals where Google would send readers to news sites, and Google would pay through advertising.

The current fight with AI is different in one important way: AI can write out an article's main points or even copy whole sentences from it, so readers might never need to visit CNN's website at all. With Google search, readers still had to click through to read the full article. This is a bigger threat to how news companies make money.

The Technical Question

Here is where it gets tricky: there is a difference between using articles to train an AI system, and using articles right now to answer a question.

Many AI companies argue that using public web content to train their systems is fair use — a legal idea that lets people use copyrighted material in limited ways without permission. But Perplexity works differently. It looks at the web right now, in real-time, when you ask it a question. It does not just rely on things it learned during training. That might make it harder to defend in court.

The paywall issue is also important for CNN's case. If Perplexity is truly copying CNN's work and showing it to people who did not pay, CNN can show the court that Perplexity caused real money damage to its business.

Other AI Companies Are Making Deals

Some AI companies have recognized this problem and struck publishing deals. OpenAI signed agreements with The Atlantic and News Corp. Anthropic made a deal with Time magazine. But even these deals did not stop lawsuits — The New York Times is suing anyway.

This lawsuit, along with the Times case against Perplexity and other cases working through the courts right now, will eventually set rules for how AI companies are allowed to use news content. Those rules could force AI companies to pay for content, or could change how they operate.

What unfolds in this lawsuit matters beyond just CNN and Perplexity. Major news companies are now testing their legal options in court rather than waiting to negotiate deals. This could push the AI industry toward formal licensing agreements — paid arrangements where AI companies buy the right to use content. That would help news companies, but might also limit what AI systems can do and increase their costs.