Technology · NEWS EXPLAINER
OpenAI plans visual ads during ChatGPT image generation
A US test is planned for later in October. OpenAI also announced measurement changes, while separate brand-suitability pilots remain in development.
OpenAI plans to test visual ads in ChatGPT during image generation later this month, starting in the United States with an initial group of advertisers. Its October 5 announcement says the ads will be clearly labeled and kept separate from the image being created.
The announcement combines a new advertising format with changes to measurement and brand-suitability evaluation. Those elements have different timetables: the image-generation test is planned, measurement integrations are expanding, and some evaluation tools remain experimental or under development.
Where the ad fits
The proposed format creates an advertising opportunity while ChatGPT is generating a requested image. OpenAI says the ad will remain distinct from that image. The announcement does not establish an exact screen position, a compulsory viewing period or a longer generation time.
OpenAI also says ads do not influence ChatGPT's answers. That is the company's stated policy; the announcement is not an independent audit of how every ad interaction will work.
Measuring a response versus measuring an effect
OpenAI is expanding integrations that connect advertising with conversion data and reporting. It names Hightouch, Tealium and LiveRamp for connections to advertisers' existing data systems, alongside attribution and other measurement partners.
Attribution links a recorded outcome to an ad interaction. Incrementality asks a different question: how much activity occurred because of the advertising? OpenAI describes its geographic incrementality work with Haus, Measured and WorkMagic as early experimentation. More measurement connections do not, by themselves, prove additional sales.
What the suitability pilots would assess
Brand suitability concerns whether an advertising environment fits a brand's requirements. OpenAI says it is developing evaluation pilots with DoubleVerify and Integral Ad Science, using controlled tests.
According to OpenAI, those pilots will not give the partners access to private user conversations. That assurance describes the proposed evaluation arrangement; it should not be read as independent certification or a complete account of every advertising partner's data practices.
The next details to watch
The announced starting scope is a US trial with an initial group of advertisers. No exact October start day, complete user eligibility list or worldwide launch date is provided. For readers, the next useful details are when the test actually begins, who encounters it and how the ad label and separation appear in practice.
