OpenAI has suspended a new kind of in-chat suggestion in ChatGPT after paying users reported the suggestions looked like ads for consumer brands. Company leaders insisted that there are no ads or ad tests running, but conceded that the experience wasn’t up to expectations and said the feature is turned off as they refine how it operates.
What Changed and Why OpenAI Suspended Suggestions
The messages in question came as recommendations to check out partner-built apps on the ChatGPT platform mid-conversation, according to Kullo. OpenAI’s chief research officer said that the team has turned off these suggestions to better protect the model’s accuracy and guard against irrelevant or commercial-sounding prompts. The company is also considering user-level controls so that people can dial down or turn off this type of guidance altogether.
OpenAI contends that the suggestions did not have any financial component and were intended to showcase tools that could assist users with tasks within ChatGPT. In other words, the intention was for app discovery, not paid placement. Still, the company admitted that anything that smacks of an ad needs to be treated with kid gloves on a trusted assistant.
User Backlash Is What Prompted the Pause
Complaints from Plus subscribers gained attention after screenshots showed unrelated brand names such as Peloton and Target in users’ chats for classes, including to ask technical questions.
The optics were jarring: customers who are paying for an A.I. service felt as though they were being sold something, and some wondered if the responses had been influenced by business relationships.
OpenAI executives responded in a series of public comments clarifying that there are no live ad tests and any potential monetization will be considered carefully. But the company was quick to pull back the feature — an implicit admission that drawing the line between helpful recommendations and advertising is fine indeed, especially when you’re mostly typing into a conversational interface where tone and timing are of the essence.
Ads or App Discovery: Navigating the Blurred Line
Assistant platforms are becoming ecosystems, not mere chats.
OpenAI will encourage developers to create apps that expand ChatGPT for shopping, productivity, fitness and more. Surfacing relevant tools in a conversation can be a win but it also runs the risk of users simply taking recommendations as endorsements — or ads — especially when mid-reply familiar retail brands appear.
Other platforms offer cautionary tales. App stores are a blend of editorial curation and paid placement, which is why the Federal Trade Commission recommends clear labeling whenever money comes into play. The signals are more subtle in a chat UI: A sentence on its own can seem authoritative, or even native, which is why disclosures and controls are even more important to build trust.
Strategy Signals and the Revenue Trade-Off
The pause also suggests a tension over strategy more broadly. OpenAI has recruited seasoned consumer operators to grow applications and business lines, and industry observers have anticipated advertising experiments would follow. There have been reports that leadership recently changed its focus to roll out core chat performance and maintain product quality before new monetization features.
The stakes are high. OpenAI has claimed ChatGPT works with more than 100M weekly active users, and Plus subscribers want that top-notch, no-ads help. While advertising might allow developer incentives or fund access at a lower cost, any misstep that creates a lack of trust could sideline growth and retention. In consumer research over the last year, people have repeatedly raised the need for transparency in AI systems and concern about commercial influence on automated advice.
What Comes Next for App Suggestions in ChatGPT
Look for OpenAI to experiment with clearer labeling, stricter relevance thresholds and explicit settings for app suggestions. Developers will keep an eye out for distribution opportunities that won’t turn off users, and regulators will monitor whether disclosures conform with new standards for conversational interfaces.
For now, the company’s position is straightforward: those suggestions that seemed like ads are gone, and any future tack it takes would strive to be more scientific, less mysterious and absolutely user-dominated. Whether that will be enough to win back the trust of skeptical subscribers will depend on execution — and how ChatGPT walks a fine line between helping discovery and the need to keep commercial noise out of chat.
