ChatGPT head says screenshots of ChatGPT ads are “either not real or not ads”; OpenAI's research chief says “anything that feels like an ad” is now turned off
Paid ChatGPT users recently reported seeing a prompt labeled “Shop for home and groceries.
The DecoderMatthias Bastian
Context & Ripple Effects
Reports from paid users about a shopping-oriented prompt put OpenAI's interface choices under scrutiny, and company leaders responded by drawing a line between product suggestions and advertising. That distinction matters because ChatGPT is becoming an assistant-shaped surface where commercial features can be mistaken for paid placement.
OpenAI turned off the experience its research chief said could feel like an ad, removing the immediate source of confusion for affected paid ChatGPT users.
The company publicly committed itself to distinguishing purported ad screenshots from actual advertising, raising the visibility of its interface-labeling decisions.
Second-order effects
Any future commercial prompt or recommendation feature faces a higher burden to be visibly differentiated from paid placement, especially in subscription tiers whose users may expect an ad-free product.
Advertisers and commerce partners cannot treat a product-suggestion surface as established inventory while OpenAI is still defining what qualifies as an ad; later ad testing on Free and Go tiers indicates tiering became part of that definition.
Third-order effects
If assistant interfaces increasingly combine answers, recommendations, and transactions, disclosure and placement rules will become a core trust constraint rather than a cosmetic design issue.
The episode points toward a segmented monetization model: commercial formats may be tested in selected access tiers while paid tiers preserve clearer limits, though the durable boundaries depend on implementation and user response.
The trend: Conversational AI is moving from subscription-only interfaces toward carefully segmented commercial monetization, with answer integrity and ad disclosure as the central design constraints.
I'm seeing lots of confusion about ads rumors in ChatGPT. There are no live tests for ads - any screenshots you've seen are either not real or not ads. If we do pursue ads, we'll take a thoughtful approach. People trust ChatGPT and anything we do will be designed to respect that.
The #keep4o movement has been warning about the degradation of ChatGPT for months (long time in the AI space) As well as the bad treatment and disturbing attitude from (some) OpenAI employees
This is a bad response from the Head of ChatGPT. Instead of owning the confusing UI, you're saying, “Trust me bro, the screenshots are either fake or you're dumb if you think it's an ad.” You lose users' trust this way, especially now that strong alternatives like Gemini exist. […
> be OpenAI customer > pay premium for that sweet AI magic > notice model acting weird > get labeled “delusional” by staff for complaining > mfw “it's just your imagination” > spot blatant ad mid-convo > point it out like “hey, that's an ad right?” > OpenAI exec: “any
@Yuchenj_UW “or not ads” they are “app links” not ads. nothing but respect for these pioneers but still, imagine being a shiteating exec who can only talk double
@nickaturley “It's not an ad if we just keep repeating that it's not an ad.” You guys literally announced a partnership with Target right before this. You're handling this very badly and people are noticing. [image]
head of ChatGPT pushes back on screenshots claiming ads in ChatGPT: “There are no live tests for ads — any screenshots you've seen are either not real or not ads.”
OpenAI is in an interesting situation here (re: ads) Best in class ads monetization (Meta) is ~$250 per year for U.S. users, which is roughly equal to the ChatGPT Plus ARPU So, you could serve ads just to free users, or even double revenue from paid users If their commerce