OpenAI has released an official statement of clarification after several social media users claimed they saw ads in ChatGPT—even on paid subscription plans. The matter initially went viral when screenshots circulated online showing the apparent brand promotions that pop up below the chatbot responses.
Viral screenshots trigger confusion
The issue first appeared on December 3, when an X user (formerly Twitter), @BenjaminDEKR, shared a screenshot where he claimed ChatGPT showed him an ad for US retail chain Target while researching Windows BitLocker. He said he was on a paid Plus subscription, which added to the confusion.
The post quickly went viral, crossing over half a million views, and ultimately forcing OpenAI officials to respond.
OpenAI: “These are not ads”
In response to the post, Daniel McAuley, a data scientist at OpenAI, explained that what was shown in the post was not an advertisement but an app integrated into ChatGPT. He said that OpenAI has been rolling out apps by partner brands since DevDay and that the suggestions were part of an ongoing pilot to help users discover app functionality more organically.
However, many users argued that unsolicited brand prompts inside unrelated chats effectively act as advertisements—even if not labelled as such.
OpenAI temporarily disables app suggestions
Addressing the concerns, OpenAI Chief Research Officer Mark Chen said the company "fell short" with respect to how those suggestions showed up. He confirmed that OpenAI has turned off such prompts while improving the model's accuracy and refining controls that will let users reduce or disable suggestions entirely.
“We are not testing Ads,” says ChatGPT head
Nick Turley, the head of ChatGPT, followed the post by adding that no ad experiments are running within the company. He added that any screenshots doing the rounds on the internet were either not real or did not represent the currently active ad tests.
Turley added that if OpenAI ever introduces ads, it would proceed with caution and in a very transparent manner to maintain user trust.
