The Pulse
OpenAI Adds Visual Ads to ChatGPT Image Generation
OpenAI says it will test a visual ad format during ChatGPT image generation in the United States later in October 2026. The announcement also adds measurement partners and reports early campaign results from three advertisers.
AI.info Team · Written with AI ·
“As advertising expands into AI-native environments, marketers need confidence that brand safety protections are durable, consistent, and built to operate at scale.”
Lidiane Jones, CEO of Integral Ad Science
OpenAI announced on October 5 that it will test a visual advertising format inside ChatGPT’s image-generation experience, alongside new measurement partnerships and pilots focused on brand suitability. The initial U.S. test is scheduled to begin later in October with a small group of advertisers.
The format is designed to show images that suggest how a product might be used or fit into someone’s life. OpenAI says the ads will carry clear labels and remain separate from the image a user is creating. The company also says advertising will not influence ChatGPT’s answers.
Ads will appear during image generation
The placement puts advertising in a different part of ChatGPT than the text-based formats OpenAI has previously described. Rather than appearing only beside a completed response, the new test will run during image generation, where product imagery could sit near a user’s creative task.
OpenAI frames the format as a way for advertisers to show products and experiences visually. Its announcement does not specify the ad’s exact layout, how often users will see one, or whether the test will include controls for skipping or dismissing it. The company says the ad will stay separate from the generated image, a distinction that matters when sponsored material appears within a tool that creates images on request.
OpenAI adds attribution partners
The announcement also expands the systems advertisers can use to measure whether ChatGPT campaigns lead to activity elsewhere. Integrations with Hightouch, Tealium and LiveRamp will let advertisers send conversion data from systems they already use. OpenAI lists AppsFlyer, Triple Whale, Adjust, DV Rockerbox, Northbeam, Branch, Singular, Kochava, Airbridge and Tenjin among partners supporting attribution, reporting or conversion-data connections.
For broader measurement, OpenAI says it is working with Fospha, Measured and INCRMNTAL. It is also exploring geographic experiments with Haus, Measured and WorkMagic to assess incremental impact—the additional results advertising produces, rather than activity that might have happened anyway. OpenAI describes this work as early-stage.
Early figures in the announcement come from the named measurement partners, not from an independently reported company-wide study. DV Rockerbox says WeightWatchers’ attributed cost per acquisition from ChatGPT Ads was 15.3% below its blended paid-search benchmark. WorkMagic reports statistically significant lift for wellness brand Dose, with 67% of incremental purchases coming from new customers, while Triple Whale says 93% of Portland Leather’s ChatGPT ad visitors were new.
Brand-suitability checks get outside reviewers
OpenAI says its placement guardrails are intended to keep ads out of emotionally vulnerable, sensitive or otherwise unsuitable conversations. Automated review, human oversight and ongoing monitoring are part of that system, and qualifying advertisers can use “Negative Phrases” to specify narrow placement exclusions tied to their own policies.
The company is developing brand-suitability evaluation pilots with Integral Ad Science and DoubleVerify. OpenAI says the partners will assess how its standards operate in controlled tests without accessing private conversations. That gives advertisers an external assessment of the ad environment, while the visual format’s first test remains limited to an initial group in the United States later this month.