OpenAI to test visual ads inside ChatGPT image generation in the U.S.
OpenAI is expanding advertising in ChatGPT with a visual format that will appear while users generate images, starting later this month in the U.S.
OpenAI is broadening its advertising business in ChatGPT, and one of the first new formats will place visual ads alongside the image-generation experience. The company said it will begin testing the format later this month in the U.S. with an initial group of advertisers. According to OpenAI, the ads could use images to show product inspiration, demonstrate how a product is used, or depict experiences associated with a service, giving advertisers a more visual way to promote themselves inside ChatGPT.
What the new format looks like
OpenAI described the experiment as a visual ad unit that runs during image generation in ChatGPT. In a post explaining the format and measurement plans, the company said the ads will be clearly labeled and kept separate from the image being created. The format is designed to let advertisers use imagery rather than text alone, which the company says gives them a more visual way to promote a product or service inside the chatbot.
The test is limited to the U.S. for now. OpenAI said it will begin with an initial group of advertisers as it evaluates how the format performs. The company did not detail how many advertisers would participate or how long the test would run.
OpenAI's stated ground rules
OpenAI said ads do not influence ChatGPT's answers. It also said the visual spots will be clearly labeled and remain separate from the image being generated, a distinction the company framed as part of the format's design rather than an afterthought.
The company has also said its safeguards are meant to keep ads out of emotionally vulnerable, sensitive, or otherwise unsuitable conversations. That constraint applies across ChatGPT's ad inventory, and OpenAI says it remains in place for the new visual format.
Why OpenAI is leaning into ads
OpenAI said ChatGPT now reaches 1.2 billion people every week. That audience includes many users who do not pay for a subscription, and the company has described advertising as another way to generate revenue from those users. The visual format is the latest step in that effort, following earlier work on ads and measurement tools.
The company's framing positions ads as a complement to subscriptions rather than a replacement. For users on free or lower-tier access, the trade-off is exposure to advertising in exchange for continued use of the service. For OpenAI, the appeal is a second revenue stream tied to the scale of its weekly audience.
A bigger push on measurement
Alongside the visual format, OpenAI is adding tools that let advertisers measure whether ChatGPT ads lead to purchases or other conversions. The company is integrating with Hightouch, Tealium, and LiveRamp for conversion data. Its attribution partners now include AppsFlyer, Adjust, Branch, Triple Whale, Kochava, and several others.
OpenAI is also working with DoubleVerify and Integral Ad Science on brand suitability. That work is particularly important for ChatGPT, the company said, because ads appear around conversations rather than on a traditional webpage. OpenAI said these independent partners will not get access to private user conversations while evaluating whether its advertising safeguards are working as intended.
- 1.2 billion — the number of people OpenAI says ChatGPT reaches each week.
- Later this month — when OpenAI says it will begin testing the visual ad format in the U.S.
- Hightouch, Tealium, and LiveRamp — conversion-data integration partners named by OpenAI.
- AppsFlyer, Adjust, Branch, Triple Whale, and Kochava — among the attribution partners OpenAI now lists.
- DoubleVerify and Integral Ad Science — the brand-suitability partners OpenAI is working with.
Brand suitability in a conversation setting
Ads in ChatGPT run alongside conversations rather than on a page of content, which changes how brand suitability has to be handled. OpenAI said it is working with DoubleVerify and Integral Ad Science on that problem. The company said its safeguards are designed to prevent ads from appearing in emotionally vulnerable, sensitive, or otherwise unsuitable conversations.
OpenAI also said the independent partners evaluating its safeguards will not get access to private user conversations. That commitment is meant to address a central tension in advertising inside a chatbot: advertisers need assurance about where their messages appear, while users expect their conversations to remain private.
What OpenAI has said in its own words
"Initially, we'll test this new ad format during image generation in ChatGPT. Ads will be clearly labeled, and remain separate from the image being created," OpenAI explained.
— OpenAI, in its explanation of the new ad format
How the test fits into OpenAI's ad plans
The visual format is an extension of OpenAI's existing advertising efforts in ChatGPT. The company has been building out both the creative side — new ways for advertisers to appear in the product — and the measurement side, so advertisers can trace whether a ChatGPT ad leads to a purchase or another conversion.
Attribution partners such as AppsFlyer, Adjust, Branch, Triple Whale, and Kochava are part of that measurement layer. Conversion-data integrations with Hightouch, Tealium, and LiveRamp address a different part of the pipeline. Together, they are meant to give advertisers more than an impression count.
OpenAI has not said whether the visual format will expand beyond the U.S., or what criteria it will use to decide whether to make it permanent. The company described the effort as a test with an initial group of advertisers.
What the test could mean for users and advertisers
For users, the practical effect of the test is that ads may appear in a part of ChatGPT that has so far been free of them: the moment of image generation. OpenAI's stated guardrails — clear labels, separation from the generated image, and exclusion from sensitive conversations — are the terms on which that change is being introduced.
For advertisers, the appeal is a visual placement inside a product OpenAI says reaches 1.2 billion people weekly, paired with measurement tools meant to connect ad exposure to conversions. The partners OpenAI has named cover conversion data, attribution, and brand suitability, which are the pieces an advertiser typically needs before committing budget to a new channel.
For OpenAI, the test is part of a broader effort to earn revenue from users who do not pay for a subscription. Whether the visual format works as an advertising product, and whether users accept ads in the image-generation experience, are questions the test itself may begin to answer once it starts later this month.
Sources
- BleepingComputer Original source
- explained Also reporting
Continue Reading
Trump Taps Clayton to Lead AI Task Force
President Trump names intelligence director Jay Clayton to head a new federal AI task force, following a voluntary self-policing accord with tech firms.
Anthropic Opens Voice Data Training Tap
Claude users can now opt in to share voice recordings for AI model training, with a separate toggle that keeps text chats private.
Google's Gemini 4 Argon Opens Narrowly
Google's frontier model Gemini 4 Argon reaches trusted cyber defenders first, with benchmark wins mixed and pricing set to rise.