OpenAI said on October 5, 2026, that ChatGPT users on its Free and Go plans will start seeing ads while the chatbot generates images. the company wrote that “Testing will begin later this month” in the United States with an initial group of advertisers, so the first ads are due before October ends and no exact launch day has been given.
As of October 11, nothing published by OpenAI or by outlets covering the announcement says the test has begun.
An ad format built for image generation
In its announcement of the format, OpenAI says it will “test this new ad format during image generation in ChatGPT for users on the Free and Go plans.” The ads are meant to help people picture how products and services could fit into their lives, which is why an image request, where the user is already describing a scene, is the first place the company is trying them.
OpenAI promises two guardrails on placement. “Ads will be clearly labeled, and remain separate from the image being created,” the company wrote, and it repeated its standing line that “advertising does not influence the answers ChatGPT provides.” The company also says conversations stay private and that users remain in control of their experience. Its pledge on privacy extends to the measurement side: brand suitability pilots, it says, let independent partners assess safeguards without reading private user conversations.
Those commitments echo the principles OpenAI published on January 16, 2026. In that earlier post on its approach to advertising, the company said ads would be tested for logged-in adults in the United States on the Free and Go tiers, that Plus, Pro, Business and Enterprise plans would stay ad-free, and that accounts known or predicted to belong to users under 18 would not see ads. It also said ads would stay away from sensitive or regulated topics such as health, mental health and politics.
The January post described how the first ads worked in text conversations: they sat at the bottom of an answer, appeared only when a sponsored product or service matched the conversation, and came with options to see why an ad was shown or to dismiss it and say why. It also said users could turn off personalization and clear the data used for ads at any time, and that user data would not be sold to advertisers. OpenAI’s October announcement says nothing on whether those same controls carry over unchanged to the image-generation format, so that remains a question for the test itself rather than something the record settles.
Measurement partners and early advertiser results
Much of the announcement is addressed to advertisers rather than users. OpenAI lists integrations with Hightouch, Tealium and LiveRamp that let advertisers send conversion data from their own systems, along with attribution partners including AppsFlyer, Triple Whale and Adjust, and full-funnel measurement partners Fospha, Measured and INCRMNTAL. Geo-based experiments to test whether ads cause purchases, run with Haus, Measured and WorkMagic, are described as being in early stages.
The company also cites early results from advertisers already on the platform. DV Rockerbox reported that WeightWatchers’ attributed cost per acquisition was 15.3% lower than its blended paid-search benchmark. WorkMagic reported statistically significant lift for Dose, with 67% of incremental purchases coming from new customers, and Triple Whale reported that 93% of Portland Leather’s ChatGPT Ads visitors were new. All three figures come from OpenAI’s own summary of partner reports, not from an independent audit, and each describes advertisers using ChatGPT’s existing ads rather than the image-generation format that has yet to launch.
Brand safety checks before the ads run
On the advertiser side, OpenAI says placement guardrails screen conversations for ad appropriateness, and that Negative Phrases are available to qualifying advertisers. It is also developing brand suitability evaluation pilots with DoubleVerify and Integral Ad Science, in which independent partners assess the safeguards “without accessing private user conversations.”
The practical change for people on the Free and Go plans is that an image request made in the United States may soon return an ad beside the picture. Paying subscribers on the higher tiers are outside the test under the January principles, and OpenAI’s October post does not say when, or whether, the format will reach other countries or other kinds of ChatGPT use.
The company’s phrase “later this month” is the only date it has committed to, and the October 31 end of the month is the outer edge of that window. Because the announcement describes testing with an initial group of advertisers and gives no start day, the first outside readout on how the image-generation ads perform cannot arrive before the test itself begins, and OpenAI has not said when it will report back on results from this format.
This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.
More from Morning Overview
- Hurricane Hunter radar shows four warning signs that a tropical cyclone is about to strengthen, a University of Miami study found
- Tropical Storm Rachel is dumping up to 12 inches on four Mexican states on its way to major hurricane strength
- The FTC says Lens.com doubled the price shoppers saw in Google ads
- Common allergy, bladder and sleep pills tied to sharply higher dementia odds
