OpenAI said on 5 October that it is adding a visual ad format to ChatGPT, shown while the model generates images. Testing begins later this month in the United States with an initial group of advertisers. The company describes the format as conveying “product inspiration, product usage, or the experiences they make possible”, and says ads “will be clearly labeled, and remain separate from the image being created”.
That labelling sentence is a partial answer to a question left open in September, when OpenAI put sponsored agents inside ChatGPT without saying how they would be marked. The post also repeats the standing claim that “advertising does not influence the answers ChatGPT provides”.
The bulk of the announcement is measurement plumbing, which is the better guide to intent. OpenAI names partners in four categories: data integration (Hightouch, Tealium, LiveRamp), click attribution (AppsFlyer, Adjust, Branch, Kochava and others), incrementality testing through geo-based experiments (Haus, Measured, WorkMagic), and full-funnel measurement. DoubleVerify and Integral Ad Science are building brand suitability pilots, which OpenAI says will work without access to private conversations.
Two things the post does not say: whether paying subscribers see these ads, and how personalisation works. Why it matters: a measurement stack is what turns an advertising experiment into an advertising business, and OpenAI puts ChatGPT at 1.2 billion people a week.
Source: OpenAI’s announcement.