ChatGPT introduces a new ad format placing product carousels on the image generation loading screen

Meta description: OpenAI will test labeled display ads below ChatGPT generated images in the US, adding measurement tools and brand safety pilots as ad revenue grows.

Tags: ChatGPT Ads, OpenAI, Display Advertising, AI Monetization, Ad Measurement

Featured image: abstract technology style, showing a generated image canvas with a separate labeled ad carousel below; no real people, no company logos, no visible text.

Background: From ChatGPT Subscriptions to an Advertising Layer

OpenAI is preparing a new test of ChatGPT display ads inside the image generation workflow, a change that matters because it extends a newer revenue channel that is already large enough to be measured in annualized dollars. OpenAI has been running ads in ChatGPT since February, and those ads are now available in more than 40 countries. The company has reached an annualized revenue run rate of $1 billion from ads alone, while ChatGPT reaches 1.2 billion people weekly. Earlier this year, only about 5 percent of users were paying for a subscription, a context that helps explain why OpenAI is testing formats that can be shown to a broader audience without changing the core chat experience.

Development: How the New Image Ad Test Will Appear

The upcoming test focuses on US users who ask ChatGPT to generate an image. During that moment, OpenAI plans to show display ads below the generated visual. Each ad will carry an advertising label and remain visually separated from the generated picture, so sponsored units do not blend into the AI output. The format will appear either as a carousel or as a row of product images. OpenAI designed these ad units to present products while they are being used or to highlight experiences associated with them, which fits the visual nature of image generation. The company is launching the test later this month with an initial group of advertisers.

Latest Progress: Measurement Tools and Brand Safety Checks

OpenAI is also expanding the measurement infrastructure around ChatGPT Ads. Advertisers can send conversion data to ChatGPT Ads through Hightouch, Tealium, and LiveRamp. Attribution providers such as AppsFlyer, Adjust, and Branch are also joining the mix, helping advertisers understand which interactions may be connected to a sale, signup, or other desired action. In parallel, OpenAI is testing geo-based experiments through Haus, Measured, and WorkMagic. These experiments compare locations where ads are shown with places where they are not, giving advertisers a way to estimate actual ad impact rather than relying only on clicks or impressions.

Brand safety is the other visible part of the test. OpenAI is running pilot programs with DoubleVerify and Integral Ad Science. Those partners evaluate the ad environment without accessing real user conversations, a design choice intended to keep advertising checks separate from private chat content. Select advertisers can also set negative keywords, a control that helps keep their ads from showing up in unsuitable conversations.

Current Stage: A Limited US Test Inside a Fast Growing Ad Business

The image ad experiment is still narrow: it is aimed at US users, it begins with an initial group of advertisers, and it sits below the image generation output rather than replacing the chatbot's main response flow. For users, the change introduces commercial content into a creative moment. For advertisers, it creates a visual ad surface tied to generative output, supported by measurement partners and brand safety pilots.

The next phase will show whether OpenAI can scale the format beyond its first advertiser group while keeping ad labeling, measurement, and brand safety clear enough for both users and buyers. The company's stated target is $100 billion in ad revenue by 2030, so the image test is a small but concrete step toward making ChatGPT a visual advertising surface as well as a conversational assistant.