WaPo inks OpenAI deal, putting journalism inside ChatGPT

The Washington Post has become the latest legacy publisher to cut a deal with OpenAI, putting its reporting directly into ChatGPT search responses. The partnership means that when users ask the chatbot questions about complex or breaking news, they might see summaries, direct quotes, and attribution links back to Post articles rather than just a synthesized blob of web-sourced information.

Post executives are framing this as a distribution play. Peter Elkins-Williams, the paper’s Head of Global Partnerships, said the team is “all in on meeting our audiences where they are,” which increasingly means inside AI interfaces instead of on a homepage or in a dedicated app. For a publication that has aggressively courted digital subscribers, that’s a notable pivot — one that acknowledges the platform wars are shifting from social media feeds to generative search.

OpenAI’s Varun Shetty points to scale as the justification, noting that over 500 million people now use ChatGPT weekly. For OpenAI, deals like this are both a defensive moat against lawsuits from publishers who’ve argued their content was scraped without permission, and a practical way to reduce the hallucination problem by anchoring responses to vetted journalism. The Post joins a roster of more than 20 publishers OpenAI has signed, spanning over 160 outlets.

What’s less clear is the economic arrangement. These deals are typically structured as licensing agreements with undisclosed terms, and neither side is talking dollars. The Post itself has been building its own AI tools, including an “Ask The Post” feature and a climate-focused chatbot, suggesting a dual strategy: license the archive to the big platforms while trying to own the direct reader relationship in-house. Whether handing your journalism to a synthetic concierge strengthens or cannibalizes that relationship is the multi-billion dollar question nobody has answered yet.