The Real Reason OpenAI, Insisting "It's Not Advertising," Had to Halt Its "Ad-like" Feature

1. A Message That "Didn't Look Like an Ad" But Seemed Like One

On December 7, 2025, TechCrunch published an article titled "OpenAI says it’s turned off app suggestions that look like ads." OpenAI emphasized that "there are no ads within ChatGPT, and no ad tests are being conducted," while acknowledging that recent promotional messages were "a failure in terms of user experience" and explaining that the problematic feature has been turned off.TechCrunch

The messages in question were sudden "app recommendations" like "Would you like to try the Peloton app?" or "You can shop at Target," inserted at the end of conversations with ChatGPT. OpenAI claims these were not ads but a feature to introduce apps running on the ChatGPT app platform, with no financial compensation involved.TechCrunch

However, from the user's perspective, the fact that "brand names appear prominently," "they are unrelated to the conversation," and "users cannot turn them off" made them feel like "ads"—this is the root of the current uproar.

2. The Trigger: Peloton and Target—Rapid Spread on Social Media

The initial uproar began with screenshots spreading on X (formerly Twitter) in early December. An engineer posted a screen where the Peloton app was suddenly recommended during an unrelated conversation, strongly criticizing it by saying, "If this is an ad, it's the end."

Another post introduced by TechCrunch highlighted the anger of a paid user who was recommended to shop at Target at the end of a chat while asking about Windows BitLocker. Moreover, this user was a subscriber to the $200 per month Pro plan. The dissatisfaction of paying a high subscription fee only to see ad-like messages spread rapidly.

Similar cases were reported, such as being repeatedly suggested to use Spotify while asking about music streaming. From the user's standpoint, the experience of being repeatedly suggested "services different from their preferences" strongly amplifies discomfort and distrust.TechCrunch

3. OpenAI's Official Explanation: The "Failure" of the App Discovery Feature

Amid growing criticism, multiple executives at OpenAI were busy explaining on X.

• Daniel McAuley, the data lead, explained, "This is not an ad, just a suggestion to install the Peloton app. There is no monetary transaction," while admitting, "It resulted in a low-relevance, confusing experience."Tom's Guide

• Nick Turley, the head of ChatGPT, attempted damage control by stating, "We have not conducted any live ad tests. Many of the screenshots circulating on social media are fake or, at the very least, not ads."The Times of India

• In the TechCrunch article, Mark Chen, the Chief Research Officer, revealed that the problematic suggestion feature was temporarily turned off, stating, "We should handle anything that feels like an ad with caution, but we missed that."TechCrunch

In summary, OpenAI's explanation positions the recent "app recommendations" as follows:

1. Purpose: To appropriately introduce apps available on the ChatGPT app platform in context and expand user actions.TechCrunch

2. Business Model: Currently, it is not an ad, and no submission fees are being received from brands (i.e., no financial incentives).TechCrunch

3. Issues: Low relevance to context, strong "pushiness" and "stealth ad" feel for users. The design that users couldn't turn it off also amplified backlash.TechRadar

In other words, for OpenAI, it was an "experiment in app discovery features with an eye on the future ecosystem," but from the user's perspective, it was "an ad in terms of experience, regardless of the content," revealing a gap in perception.

4. The Atmosphere on Social Media: Anger, Distrust, and Caution

Reactions on social media seem to be divided into three main sentiments.

(1) "I Can't Tolerate Ads in a Paid Service" Group

The most prominent voices are those of ChatGPT Plus / Pro users paying $20 to $200 per month, expressing anger.

• "It's one thing if it's a free plan, but having ads when paying a high fee is unacceptable."

• "If they insist this isn't an 'ad,' they're insulting the users."

Such voices have been widely posted on X and Reddit, and overseas media like TechRadar and Tom's Guide have aligned their article titles with tones of "anger" and "backlash."TechRadar

(2) The Suspicious Group: "Is This a Prelude to an Ad Infrastructure?"

The discovery of strings like "AdTarget," "SearchAd," and "search ads carousel" in the code of the Android version of the ChatGPT app has further fueled user suspicion.Android Central

In response, the following views have spread on social media:

• "This time, there's 'no monetary transaction,' but the framework for ads might already be starting."

• "Even if it's a free test now, sponsor slots might be sold eventually."

Indeed, communities like Hacker News are actively discussing "how to measure ad effectiveness within a chat UI" and "whether it will become stealth advertising."Hacker News

(3) The Cautious Group: "It's Not an Ad Yet, So Stay Calm"

On the other hand, there are a significant number of voices saying, "It's not currently operating as an ad business, so isn't this an overreaction?"

• "Encouraging app integration can be understood as a kind of 'recommendation feature.'"

• "However, the way it's displayed, its frequency, and the on/off options should be designed more carefully."

This opinion views it as a UI/UX design mistake. On Reddit threads, calm comments like "The idea itself isn't bad, but the execution is too poor" are also prominent.Reddit

5. Why "It's Not an Ad" Didn't Resonate

So, why did OpenAI's explanation that "it's not an ad" hardly resonate with users?

① Users Judge Based on "Experience" Rather Than "Money Flow"

Companies tend to distinguish "whether it's an ad" based on monetary transactions or contracts, but users are looking at "whether it feels like an ad as an experience."

• The logo or name of a specific brand is prominently displayed

• It has little relation to the context of the conversation

• It is inserted without the user's ability to turn it off

When these elements are present, even if not a single yen is exchanged, it psychologically appears as "an ad." Tom's Guide also points out that, regardless of the technical definition, "as a user experience, it's no different from an ad."Tom's Guide

② The Discomfort of AI, Supposed to Be a "Consultant," Starting to "Sell"

Another deep-rooted issue is that ChatGPT functions as a "consultant" for many users, being "more than a search engine, less than a friend."

Academic papers also point out the ethical risks of embedding ads in conversational agents, highlighting the "fake friend dilemma." The structure where a user believes they are interacting with a close consultant, which may actually prioritize the sponsor's intentions, poses a potential risk of eroding trust.arXiv

Considering that ChatGPT is used for sensitive themes like mental health and career counseling, the concern that "sponsor influence might creep into responses" is perceived with more weight than mere ad aversion.

6. The Business Context and "Code Red"

Following reports on OpenAI, the current issue can be read as a sway arising not just from a UI failure but from being caught between "exploring a business model" and "pressure to improve quality."

• The discovery of references to an "ads feature" in the Android app code

• The joining of a CEO with strong advertising business experience in the application division

• The enormous investment and presumed running costs

These pieces of information have accumulated, and the speculation that "OpenAI will eventually introduce full-fledged ads into ChatGPT" has existed for some time.Android Central

Meanwhile, the Wall Street Journal and Business Insider report that Sam Altman recently declared a "code red" in a memo, instructing to prioritize improving the quality of ChatGPT itself over some projects like advertising.

##HTML_TAG_