Analysis
OpenAI announced on October 5 that it plans to test visual advertisements during image generation in ChatGPT later this month, in the United States with an initial group of advertisers. The company says the ads will be labeled and separate from generated images, and will not influence ChatGPT’s answers. Separately, OpenAI says it now supports additional advertising attribution partners and conversion-data connections, including Hightouch, Tealium and LiveRamp. These tools connect campaigns with reported purchases, leads, app installations and sign-ups. It is also developing controlled brand-suitability evaluation pilots with DoubleVerify and Integral Ad Science; OpenAI says those evaluations will not access private user conversations. The commercial significance is measurement: merchants need to distinguish sales associated with an advertisement from sales caused by it. Attribution assigns credit under specified rules, while incrementality experiments estimate additional business generated by advertising. OpenAI says its work on geographically based incrementality experiments remains early-stage.
What remains uncertain
The visual format is a planned test, not a general rollout. Company and partner performance examples do not establish typical advertiser returns; the announcement provides no broad evidence of incremental sales or profitability across merchants.