AI NewsGo-to-marketReported
OpenAI engineer told colleagues users won't click links, and a court filing puts Bing Chat click-through rates 87 to 93 percent below Bing Search
Search Engine Journal reports a news-publisher court brief quotes an OpenAI engineer saying users will not click on links no matter how prominent, and cites Microsoft data showing click-through rates in Bing Chat were 87 to 93 percent below Bing Web Search for New York Times sites.

Image: Search Engine Journal
Why it mattersFor any site whose traffic depends on being cited by an answer engine, the reader has a first concrete cross-product comparison of what a citation is now worth in clicks, and it is a small fraction of a normal search click.
Search Engine Journal has published an SEO story that puts a real number on something publishers have said for two years: a link inside an AI chat answer is worth far fewer clicks than the same link inside a search result. The source is a publishers' court filing, so the figures are one side's argument in an unresolved case, and worth reading in that context.
Matt Southern at Search Engine Journal reports that a combined summary judgment brief filed on 17 September by news publishers suing OpenAI and Microsoft in federal court in New York quotes an OpenAI engineer as writing, in February 2023, that "no matter how prominently we show the links, users won't click". The same brief quotes Nick Turley, head of ChatGPT, saying there is "no good reason to click" on a source link once ChatGPT has answered the question. Both quotes were carried into the SEJ piece from the publishers' brief and from The New York Times' own reporting on it.
What the Microsoft data says
The publishers' filing also cites Microsoft's own click data, comparing the click-through rate on links to publisher sites in Bing Chat with the click-through rate on the same publishers in Bing Web Search. Southern reports the ranges the brief gives: 87 to 93 percent lower in Bing Chat for New York Times sites, 83 to 91 percent lower for the Daily News plaintiffs' sites, and 51 to 94 percent lower for Ziff Davis sites. The brief defines click-through rate as clicks on a URL as a share of the times that URL appeared, either in a search results page or in a chatbot's response. That measures behaviour on the results surface, and it is a different thing from total traffic or total visits.
The brief does not say when Microsoft collected the data, how many queries or impressions it covered, what the low and high ends of each range stand for, or how the sites were selected. Microsoft renamed Bing Chat to Copilot in November 2023, and nothing in the passage ties the figures to the current Copilot.
What OpenAI and Microsoft are arguing back
OpenAI and Microsoft filed their own summary judgment motions on 4 September and argue their use of publisher content is fair use. A Microsoft spokesperson told TheWrap the company's filings explain why Copilot "is not a substitute for publishers' journalism". OpenAI has not commented on the specific engineer quote in the story reported by SEJ. Judge Sidney H. Stein's schedule sets 23 October for oppositions and 20 November for replies, so the responses to these quotes will appear in the next round of filings.
The numbers matter for a team building software or content that expects to be found through an answer engine. Bing and Google both give creators a citation count in their AI reports and do not give a click count. Microsoft's Citation Share in Bing Webmaster Tools notes it does not represent traffic share, and Google Search Console's generative AI report shows impressions with no clicks. A citation is easier to earn than a click, and the publishers' filing is the first cross-product number that says how much easier: for the sites the brief covers, an AI answer citation drives roughly a tenth of the clicks a Bing search result would.
Source
Primary source: Publishers' combined summary judgment brief, 17 September 2026 (PDF).
This item was written by an AI system from the linked source. Reveneau is responsible for what it publishes.
Get AI News in your inbox
New developer tools, model and agent releases, and how teams are actually using them to release software. Short, and only when there is something worth reading.

