Guardian inks deal to put journalism inside ChatGPT

The Guardian and OpenAI announced a strategic partnership Friday that will pipe the British newspaper’s reporting directly into ChatGPT, giving the chatbot’s 300 million weekly users access to extended summaries of Guardian articles. The deal includes attribution and links back to source material — a transparency measure that has become standard in these arrangements as publishers try to thread the needle between embracing AI and protecting their content.

But this isn’t a one-way street. Guardian Media Group will also roll out ChatGPT Enterprise across its own operations, with plans to develop new tools for both readers and the business side. It’s a notable move for a publication that just a year ago published a framework emphasizing human oversight of generative AI. The internal deployment suggests the company sees genuine utility beyond the licensing revenue.

Keith Underwood, the Guardian’s CFO and COO, framed the deal around intellectual property rights — language that matters. After years of media companies watching their work get scraped for training data without compensation, the tone has shifted. “This new partnership with OpenAI reflects the intellectual property rights and value associated with our award-winning journalism,” Underwood said. OpenAI COO Brad Lightcap called it part of a “strategy to help publishers and audiences benefit from advanced AI technology.”

The Guardian joins a growing list of publishers — including Axel Springer, Le Monde, and the Financial Times — that have cut deals with OpenAI rather than fight in court. But each partnership raises the same thorny question: when a chatbot summarizes the news, how many users actually click through to the original? The attribution links are there, but the economic equation for publishers remains unproven.