OpenAI Clarifies It Isn’t Testing Ads on ChatGPT Despite User Claims

OpenAI officials have dismissed claims of ads appearing on paid tiers of ChatGPT. The controversy initially stemmed from screenshots shared by users on social media recently, highlighting ads appearing below prompt responses in an unrelated conversation. This happened even if they were enrolled in a paid subscription to the AI chatbot. Following the backlash, an OpenAI official claimed that what screenshots of users saw on social media was either “not ads or not real”.

Ads on ChatGPT

In a post on X, OpenAI's Chief Research Officer Mark Chen wrote that anything that feels like an advertisement within the AI chatbot “needs to be handled with care”. The official admitted that here, the company “fell short”.

The most notable instance of ads appearing on ChatGPT surfaced on December 3, when an X post by user @BenjaminDEKR said that they were researching Windows BitLocker on ChatGPT, but they came across ads. Below the response, the AI chatbot showed an ad about shopping for home and groceries at the US retail store Target, which was unrelated to the conversation.

“I'm in ChatGPT (paid Plus subscription), asking about Windows BitLocker, and it's showing me ADS TO SHOP AT TARGET”, the post read. The social media post garnered over half a million views, prompting an initial response from Daniel McAuley, a Data Scientist at OpenAI.

The official clarified that what the user saw was not an ad, but an app integrated into ChatGPT. “We've launched apps from some of our pilot partners since DevDay, including Target, and have been working to make the discovery mechanism for apps more organic inside chatgpt”, the response read.

However, the user responded by claiming that when brands inject themselves into an unrelated chat and encourage users to shop at their store are essentially ads.

Responding to the user, OpenAI's Mark Chen clarified that the company has turned off such suggestions while the model's precision is improved. Further, OpenAI is also said to be incorporating better controls to enable the user to reduce its frequency or turn it off altogether, if they wish to.

In a subsequent X post, Nick Turely, Head of ChatGPT, confirmed that it is not running any live tests for ads, and any screenshots that users may be seeing are either not ads or not real.

“If we do pursue ads, we'll take a thoughtful approach. People trust ChatGPT and anything we do will be designed to respect that,” Turley added.