OpenAI Disables App Suggestions That Resemble Ads
▼ Summary
– OpenAI executives responded to paying user complaints about seeing promotional messages for companies like Peloton and Target within ChatGPT.
– The company stated these were tests for showing apps from its platform, not ads, and had “no financial component,” a claim some users disputed.
– Executives, including head Nick Turley, insisted there are no live tests for ads and promised a thoughtful approach if advertising is ever pursued.
– Chief Research Officer Mark Chen apologized, acknowledged the company “fell short,” and said these suggestions have been turned off while improvements are made.
– A recent internal memo reportedly declared a “code red,” prioritizing ChatGPT’s quality and pushing back other products, including advertising development.
OpenAI has taken action to disable certain suggestions within ChatGPT after user complaints that the prompts resembled unwanted advertisements. The company maintains it is not currently running ads, but acknowledges the promotional messages missed the mark for its user experience. Chief Research Officer Mark Chen stated the company “fell short” and has turned off the problematic suggestions while working to improve the model’s precision. This move follows significant feedback from paying subscribers who expressed frustration over seeing prompts for brands like Peloton and Target during their conversations with the AI.
Users shared screenshots on social media showing the assistant, in the middle of answering a technical question about Windows BitLocker, suggesting the user might want to shop at Target. One subscriber posted the image with the caption, “Yeah, screw this. Lose all your users,” highlighting the depth of the negative reaction. In response to the growing controversy, executives took to public forums to clarify their position. ChatGPT head Nick Turley addressed what he called “confusion about ads rumors,” insisting that no live tests for advertisements are occurring and that any concerning screenshots are either fabricated or not actual ads.
He added that if OpenAI ever decides to pursue an advertising model, it would adopt a “thoughtful approach” designed to respect user trust. Chen’s response struck a more conciliatory tone, directly agreeing that anything perceived as an ad must be handled carefully. Beyond disabling the feature, he noted the team is exploring better user controls, potentially allowing individuals to reduce or turn off such suggestions entirely if they do not find them valuable. The company explained the prompts were part of a test to highlight apps built on its recently announced platform, emphasizing there was “no financial component” to those specific suggestions.
This incident occurs against a backdrop of internal shifts at OpenAI. Earlier this year, the company brought on Fidji Simo, a former executive from Instacart and Facebook, as CEO of Applications. Her hiring led to widespread industry speculation that she would spearhead the development of an advertising business. However, a recent report suggests a change in priorities. According to the Wall Street Journal, OpenAI CEO Sam Altman issued a memo declaring a “code red,” focusing company efforts on enhancing the core quality and reliability of ChatGPT. This strategic pivot has reportedly delayed other product initiatives, including any formal plans for advertising, as the company works to rebuild and maintain user confidence in its flagship service.
(Source: TechCrunch)
