OpenAI insists ChatGPT ‘ads’ weren’t ads

OpenAI is trying to reassure paying users after ChatGPT surfaced what looked like ads from brands like Target and Peloton – and said those prompts weren’t ads at all.

What happened. Paying ChatGPT users shared screenshots of promotional-style prompts, including a message urging them to “Shop for home and groceries. Connect Target.”

• In response, OpenAI said those prompts were recommendations for apps built on the ChatGPT platform, with “no financial component.”
• Users saw a brand logo and a call to action – and drew the obvious conclusion: these were ads.
• Users were not pleased. One response: “Bruhhh… Don’t insult your paying users.”

What OpenAI is saying. OpenAI leaders moved quickly to clarify:

• “No ads.” ChatGPT head Nick Turley said there are “no live tests for ads” and claimed the screenshots were “either not real or not ads.”
• Not an ad, but… Chief research officer Mark Chen struck a more apologetic note, admitting OpenAI “fell short” and that anything that feels like an ad must be “handled with care.”
• Feature off. Chen said OpenAI has disabled these app suggestions while it sharpens the model’s precision and adds controls so users can dial them down or turn them off.

Ads on hold. Despite all the discussion about ChatGPT ads, OpenAI paused its advertising plans. CEO Sam Altman declared a companywide “code red” to rapidly improve ChatGPT’s quality, according to an internal memo. That push is delaying several initiatives, including advertising.

Why we care. We thought we were seeing the first ChatGPT ads, but that wasn’t the case. So for now, we wait until ChatGPT comes out of its “code red” status with its new and improved product experience.

More coverage. See Techmeme