What the new ads look like
On October 5 at Advertising Week New York, OpenAI announced a new ChatGPT ad format: when users ask ChatGPT to generate images, visual display ads will appear alongside. They launch later this month in the US only, with an initial test group of advertisers. According to OpenAI, the ads can use images to show product inspiration, how a product is used, or the experience a service offers.
OpenAI's stated rules: ads are clearly labeled and kept separate from the image being created; they don't influence ChatGPT's answers, and advertisers can't shape or rank responses; conversations aren't shared with advertisers; and no ads appear in emotionally vulnerable or sensitive conversations. DoubleVerify and Integral Ad Science will evaluate brand safety. On measurement, OpenAI added more than a dozen attribution and conversion partners, including AppsFlyer, Triple Whale and Adjust, plus geo-based incrementality experiments with Haus and others.
From text ads to image ads
ChatGPT introduced ads for free and low-cost plans earlier this year and expanded them to India in August. The first format was a single sponsored unit at the bottom of an answer, charged per click. Image generation is a natural inspiration moment, when users are already picturing what they want, so a relevant product image there is closer to shopping guidance than a link at the end of an answer.
What it means
For advertisers, ChatGPT's 1.2 billion weekly users now have a visual entry point, which becomes a new channel once it rolls out globally. For users, "separate from generated content" and "no influence on answers" are still promises that haven't been tested in a live product at scale. TechCrunch also notes the ads could push more casual users to pay for a subscription to get rid of them.