OpenAI has switched off a set of promotional app messages inside ChatGPT after users complained they looked and felt like ads — including for paying subscribers. The company says the experiment was meant to help people discover integrated apps, but acknowledges it mishandled how those suggestions showed up.
In a post on X, OpenAI chief research officer Mark Chen said the team has “turned off this kind of suggestion” while it works on improvements, adding that anything that feels like an ad “needs to be handled with care” and that the company “fell short” this time.
ChatGPT Users Spot Shopping Prompts in Random Chats
The backlash kicked off after ChatGPT users — including some on paid Pro and Plus plans — started sharing screenshots of shopping-style prompts appearing underneath unrelated conversations.
People chatting about things like xAI or Windows BitLocker were suddenly seeing call-to-action boxes nudging them to “find a fitness class” or “shop for home and groceries.” Those boxes linked directly to Peloton and Target apps that run inside ChatGPT.
To many, it looked like the first wave of in-app ads had quietly arrived. The timing didn’t help: the prompts appeared below conversations, in a UI slot that felt very much like sponsored units in other apps.
OpenAI: “Not Ads,” Just App Suggestions
OpenAI insists this wasn’t an ad rollout.
A company spokesperson told TechCrunch the messages were part of an experiment to surface partner apps more prominently, pointing back to an October announcement that ChatGPT would start suggesting apps when they’re relevant to a conversation.
OpenAI data engineer Daniel McAuley responded to one user on X, stressing that there was “no financial component” to the suggestions and arguing that they weren’t ads. But he also admitted that the lack of relevance made the experience “bad” and confusing. In a separate reply, he said the idea was to encourage organic discovery of apps inside ChatGPT so users wouldn’t have to jump out to other services.
That nuance was mostly lost on people who suddenly found retail prompts sitting under technical or unrelated chats.
Chen later followed up publicly, acknowledging the misstep and promising better controls so people can reduce or switch off similar suggestions in the future.
Ads or Not, the Monetization Pressure Is Real
The intensity of the reaction says as much about user expectations as it does about the feature itself.
ChatGPT has been free of traditional advertising since launch. For a lot of people, especially paying subscribers, the appearance of shopping-style prompts set off alarm bells that the interface was about to go the way of social feeds and search results.
There’s also the bigger context: OpenAI is under heavy pressure to turn ChatGPT’s huge audience into durable profit. The company has reportedly hit around $12 billion in annualized revenue as of summer 2025. At the same time, internal projections suggest OpenAI expects to burn roughly $115 billion through 2029, and it has signaled plans to spend more than $1 trillion this decade on compute and infrastructure as it chases superintelligent AI.
Most of today’s income comes from API usage and ChatGPT subscriptions. But according to a Financial Times report, only about 5 percent of ChatGPT’s roughly 800 million users actually pay for the service, leaving a huge gap between usage and direct monetization.
Seen through that lens, experiments that push people toward shopping, partner apps, or other revenue-adjacent flows are not surprising — even if OpenAI insists this particular test wasn’t about ad dollars.
“No Live Tests for Ads,” but Future Ads Aren’t Off the Table
Nick Turley, who leads ChatGPT, publicly pushed back on the idea that OpenAI has already rolled out advertising.
Responding to viral screenshots, Turley said there are “no live tests for ads” inside ChatGPT and claimed that any screenshots circulating are either not real or not actually ads. He did not specify which of the posted images might be fabricated versus simply misinterpreted experiments.
Turley has, however, been candid in the past about keeping the door open. In an August interview on Decoder, he said he wouldn’t rule out bringing ads to ChatGPT someday, but stressed that any move in that direction would need to be “very thoughtful and tasteful.”
OpenAI CEO Sam Altman has made similar comments, saying he’s “not totally against” ads in ChatGPT and even praising how Instagram integrates advertising into its feed.
For now, OpenAI is pausing anything that looks remotely like an ad inside ChatGPT as it deals with what’s been described internally as a “code red” moment: rising competition from Anthropic, Google, and others, plus questions about how to monetize AI assistants without annoying the very users that made them popular. Google, for instance, is already testing ad formats inside its AI-powered Search mode, while also delaying Gemini chatbot ads until at least 2026 after watching rivals stumble.
OpenAI says it wants to improve the relevance and controls around app suggestions before anything similar returns. The bigger question is whether this episode ends up as a one-off misfire — or the first glimpse of what ChatGPT looks like once the ad business finally arrives.