OpenAI tests ads in ChatGPT

OpenAI has started testing ads for logged-in adult Free and Go users of ChatGPT in the US. Ads will not be shown to Plus, Pro, Business, Enterprise, Education plan users.

Ads do not influence responses, and chat content is not shared with advertisers. Displayed ads are matched based on conversation topics, past chats, and ad interactions, and are always labeled sponsored, visually separated from responses.

Users can hide ads and leave feedback, check why a particular ad is being shown, delete ad data all at once, and change ad personalization settings. On the Free plan, users are also given the option to turn off ads in exchange for a reduced number of daily free messages.

Safeguards have also been added. Ads are not shown to accounts estimated to be under 18, and ads are not displayed around sensitive or regulated topics such as health, mental health, and politics. Advertisers cannot access chats, memory, or personal information, and only receive aggregated view and click counts.

In the March 26, 2026 update, OpenAI stated that initial pilot results were positive. There was no impact on consumer trust metrics, the ad dismissal rate was low, and ad relevance continues to be improved based on feedback. Accordingly, the pilot is expanding beyond the US to Canada, Australia, and New Zealand.