User backlash forces OpenAI to disable ChatGPT AI app suggestions
The core problem centered on the AI chatbot’s user experience. For many users, particularly those paying for a subscription, ChatGPT represented a clean, ad-free environment. When these branded suggestions appeared—for instance, recommending the Zillow app during a house-buying discussion—users perceived them as the platform’s first official, unannounced ads. Screenshots quickly circulated online, fueling the controversy and eroding trust among the user base.
The situation sparked an internal debate among OpenAI executives. Vice President Nick Turley quickly denied that the company was running “live tests for ads.” He insisted that the suggestions were simply misidentified. He assured users that the company valued their trust and would approach any future monetization thoughtfully.
Users mistook the feature for unwanted ads
However, users argued that the company’s intent did not matter. At the end, the people perception of advertising was what counted. Chief Research Officer Marc Chen took a more conciliatory approach . He acknowledged the user sentiment, agreeing that anything that felt like an ad needed to be handled with care. In response, OpenAI shut down the app recommendations feature on ChatGPT. Chen also announced plans to both improve the model’s precision in delivering suggestions and explore options for introducing user controls. These potential options should allow people to reduce or turn off the recommendations entirely.
This quick retreat suggests OpenAI recognizes the sensitive balance required when introducing commercial elements to a platform built on simplicity. The controversy coincided with reports of CEO Sam Altman issuing an internal “code red.” The latter aimed to prioritize improving the model’s core quality, temporarily pushing back other ambitious projects—including early plans for advertising.
The message is clear: while some form of monetization is likely inevitable, OpenAI must find a transparent way to present commercial content. If or when app suggestions return, they must look and feel fundamentally different from the current implementation. Otherwise, the company will run into the same user backlash.
