OpenAI staff say people won't click links in AI answers
The line was found buried in 92 pages of the New York Times case against OpenAI.
An OpenAI engineer is quoted in a newly unsealed court filing saying that people do not click the links in AI answers, however obvious those links are made.
The SEO Natzir found the line in the New York Times case against OpenAI and posted it, calling it a traffic story buried in 92 pages. Gagan Ghotra shared it on to a wider audience with the reaction "JUST WOOOWWW".
The quote itself is short. "No matter how prominently we show the links, users won't click," the engineer is quoted as saying in the filing.
No matter how prominently we show the links, users won't click.
Why this matters for your site
If you have been told to chase citations in ChatGPT the way you chase blue links in Google, this is a company insider saying the click behaviour is different. Being named in an AI answer is not the same as someone landing on your page. Nobody outside the case knows how broadly that engineer's view is held inside OpenAI, and the filing quote is all we have so far.
It lines up with what we covered when research said AI answers cut clicks without making users any happier. Different companies, same pattern. The answer gets read, the source does not get visited.
What to do about it
Do not rebuild your whole plan around a single quote from a court document. Do treat AI mentions as brand exposure rather than as traffic you can count on.
One practical step. Open your analytics and look at how many visits came from chatgpt.com or other AI tools last month, then compare that to search. If AI referrals are a rounding error, stop worrying about them and keep writing the pages that answer real questions for real customers, because those are the pages that get quoted in the first place.
And if your business depends on people reaching your site, make sure the things they need most, your prices, your booking form, your phone number, are worth the click when someone does make it.