OpenAI faces backlash after ChatGPT suggests an app that looked like an ad
What happens when your AI assistant unexpectedly starts recommending apps mid-conversation—especially when you’re paying hundreds per month for an ad-free experience? That’s exactly what stirred up controversy for OpenAI this week.
A ChatGPT user found themselves on the receiving end of a Peloton app recommendation during an unrelated chat, sparking widespread assumptions that ads had quietly made their way into OpenAI’s chatbot—even for premium subscribers. The outcry was quick and loud, but OpenAI clarified that the suggestion wasn’t a paid ad. Instead, it was an experimental feature aimed at helping users discover apps through organic conversation. The problem? It didn’t feel organic at all.
The conversation began when Yuchen Jin, co-founder of AI startup Hyperbolic, posted a screenshot on X (formerly Twitter) showing ChatGPT nudging them to connect the Peloton app. The post quickly went viral, racking up around 462,000 views. What really upset people was the fact that Jin was using ChatGPT’s $200-per-month Pro Plan—a price tag that usually implies freedom from anything even resembling an ad.
As the post gathered traction, others chimed in with similar concerns. One user claimed they couldn’t get ChatGPT to stop suggesting Spotify—even though they were firmly in the Apple Music camp. These users argued that paid platforms should never blur the line between experience and advertisement.
Daniel McAuley, OpenAI’s data lead for ChatGPT, stepped in to clarify the situation on X, assuring users that “this is not an ad” and “there’s no financial component involved.” He described it as merely a test of ChatGPT’s new app suggestion functionality—and admitted that the feature’s “lack of relevancy” made for a “bad and confusing experience.” According to McAuley, the team was actively revising the user experience to make it smarter and more context-aware.
OpenAI later confirmed to TechCrunch that what users saw was part of internal testing to integrate app recommendations into conversations. The company referenced its October announcement about ChatGPT’s expanding app ecosystem, which promised that apps would “fit naturally into user conversations.” As the official description put it: “You can discover apps when ChatGPT suggests one at the right time, or by calling them by name. Apps respond to natural language and include interactive tools right inside your chat.”
But here’s where it gets controversial—the supposed “natural fit” seemed anything but. Jin’s conversation had nothing to do with workouts, fitness, or Peloton. Instead, he was talking about a podcast featuring Elon Musk and xAI. The random suggestion felt forced, irrelevant, and intrusive.
Even if the app suggestion had aligned with the topic, some users argued it would still feel commercial. After all, recommending a subscription-based app inside a premium AI chat could easily resemble stealth advertising. Currently, users can’t disable these suggestions, adding to concerns about losing control of their in-chat experience.
This incident might have deeper implications for OpenAI’s larger vision. The company aims to turn ChatGPT into a central hub for app interaction—a potential replacement for mobile app stores. But if users grow frustrated or feel marketed to, that goal could backfire. People may migrate toward competing AI platforms that promise ad-free simplicity.
At present, ChatGPT’s app integrations remain in pilot testing and are available only to logged-in users outside the EU, U.K., and Switzerland. OpenAI has already partnered with big names like Booking.com, Canva, Coursera, Figma, Expedia, and Zillow, among others.
The story, originally reported by veteran TechCrunch journalist Sarah Perez, highlights how even small UX experiments can shake user trust when transparency isn’t crystal clear. After all, nobody wants their helpful AI companion to start sounding like an app salesman.
But what do you think? Should AI assistants be allowed to suggest apps—even when they’re not paid ads—or do such ‘recommendations’ cross a line into unwanted marketing? Share your thoughts in the comments—this debate’s just getting started.
