ChatGPT’s advertising business now has a documented playbook, and it looks familiar. Testing by 404 Media this week confirmed that advertisers on the platform can target users two ways: by what they are typing about right now, and by who they are. Advertisers can upload lists of email addresses, phone numbers or Android advertising IDs — raw or hashed — and OpenAI matches them against its user base of roughly one billion monthly users. A list needs at least 25,000 matched users to target an audience; smaller lists can only be used to exclude people.
The second control is the newer one. Advertisers write “context hints” — described in OpenAI’s own ad product documentation as a request to “describe the conversations, topics, or keywords where your products or services may be relevant” — which the company says “guide matching but aren’t exact-match targeting rules.” In practice, the current chat shapes which ad appears, and ads now run for Free and Go tier users, including alongside image generation results. Turning off “ads personalization” in settings does not stop ads; the setting itself notes that ads “may still be based on your current chat.” As one ad-tech observer put it, the rollout is “boiling the frog.”
What buyers should ask before spending
Customer-list targeting is standard practice across the advertising industry — the novel part is that it is now layered onto a chatbot whose users may assume their prompts are private, and which 404 Media has previously shown can be reviewed by humans. OpenAI says advertisers never see conversations, cannot tell which individuals matched their lists, and that uploaded files are deleted within roughly 24 hours. It did not say whether it hashes raw uploads before processing — a step Meta takes. Media buyers should get that answer in writing, confirm how list-match interacts with the personalization toggle, and treat the ChatGPT Ad Library as mandatory pre-flight reading.
Why this matters now
Ads launched in the United States in February 2026 and the programme has since expanded to new territories, with visual formats joining text ads inside conversations. That makes ChatGPT one of the fastest-scaling new ad surfaces in the market — and one where the targeting inputs include material users typed in confidence. For brands, the opportunity is real: intent-rich, conversational context at enormous scale. The diligence bar should be set accordingly, because the first advertiser scandal on this surface will not be about performance. It will be about what the platform knew.
