Brand Safety on ChatGPT Ads: What OpenAI Controls, and What You Cannot See
By The InPromptAds Team 7 min read
Last verified 15 August 2026. Version 1.0. Next scheduled review 15 September 2026.
Brand safety is one of the three objections that stops a ChatGPT ads deal, and it is the one with the least published detail. Here is the whole documented position, followed by an honest account of what is missing from it.
What OpenAI states
"OpenAI's policy is to place ads only near chats that are safe, appropriate, and aligned with user trust and brand safety." Safeguards are designed to prevent ads appearing in sensitive user contexts or brand-unsafe environments, including the contexts named in the ad policies.
The named exclusions from ad policies v1.3, updated 15 July 2026: no ads run beside conversations about personal health, mental health or politics, or in what the policy calls sensitive user journeys.
Separately, ads are not shown to users on Plus, Pro or any Business plan, or to accounts identified as belonging to users under 18, based on account-level age information and age prediction where available.
Logged-out ads are limited to content appropriate for all audiences, with mature content restricted to logged-in users.
The structural difference from every other channel
In display, video and social, brand safety is a control surface. You upload exclusion lists, set content categories, apply a third-party verification vendor, and receive placement reports you can audit after the fact.
None of that exists here.
| Control you have elsewhere | On ChatGPT Ads |
|---|---|
| Keyword or topic exclusion lists | Not available |
| Publisher or placement blocklists | Not applicable, there are no publishers |
| Third-party verification (IAS, DoubleVerify, MOAT) | Not available |
| Post-campaign adjacency reporting | Not available |
| Content category tiers | Not available |
| Manual placement review | Not possible, advertisers do not see conversations |
Brand safety on this channel is a platform guarantee, not a configurable control. OpenAI decides what is safe and applies it uniformly, and you either accept that or you do not buy.
That is a real answer, and for many advertisers it is a perfectly good one. It is not the answer a regulated brand's compliance team is set up to process, because their process assumes an artifact they can review.
The verification problem, stated fairly
You cannot audit adjacency because you cannot see conversations. Advertisers receive aggregated performance information, and the conversation an ad appeared beside is not part of it.
There is a legitimate reason for that, and it is not evasion: the conversations are private, they frequently contain personal information, and a platform that handed advertisers conversation-level context would be making a much worse privacy trade than the one it has made.
So the tension is real and it does not have a clean resolution. Advertiser verifiability and user privacy are in genuine conflict here, and OpenAI has resolved it in favour of privacy. An advertiser who requires adjacency proof is asking for something that would require the platform to reverse that choice.
Anyone telling you they can provide placement-level brand safety reporting on ChatGPT ads is describing something the platform does not expose.
The personalisation dimension nobody discusses
One line in OpenAI's advertiser documentation deserves more attention than it has had.
Ad selection considers the context and intent of the current conversation and, when ads personalisation is enabled, select signals from a user's broader ChatGPT experience.
So targeting is not purely conversational. Where a user has personalisation on, signals from their wider history feed selection.
For brand safety this cuts both ways and both are worth saying. It plausibly improves relevance, which reduces jarring adjacency. It also means the context your ad is matched against is partly invisible even in principle, not just unreported. You cannot reason about adjacency from the conversation alone, because the conversation is not the only input.
What to tell a compliance team
Not reassurance. The four things they will ask, answered honestly.
Where can our ad appear? Below ChatGPT responses, in conversations OpenAI's safeguards classify as safe and appropriate, excluding personal health, mental health, political conversations and sensitive user journeys.
Can we exclude topics ourselves? No. There is no advertiser-side exclusion control.
Can we verify where it appeared? No. There is no adjacency reporting and no third-party verification integration, because conversations are private.
What happens if it appears somewhere unacceptable? There is no documented advertiser-side incident process for adjacency specifically. Ad-level issues surface through the Ads tab status and the re-review path, and general support runs through OpenAI's advertiser support channel.
If a compliance team cannot approve a channel on those terms, that is a reasonable position and the honest conclusion is to wait rather than to argue them round.
What we cannot tell you
How the safeguards are implemented or how well they work. OpenAI states the policy and the intent. It does not publish a method, an error rate, or an audit.
What "sensitive user journeys" means operationally. The policy names the concept without defining the boundary, and the boundary determines both your inventory and your exposure.
Whether adjacency controls or reporting are coming. No published roadmap.
Whether any brand has had a documented adjacency incident here. We have found none reported. Absence of reports is not evidence of absence, particularly in a channel where advertisers cannot see placements.
How personalisation signals affect adjacency in practice. Documented as an input, undocumented in behaviour.
Quick answers
How does brand safety work on ChatGPT ads? OpenAI states its policy is to place ads only near chats that are safe, appropriate and aligned with user trust and brand safety, with safeguards designed to prevent ads appearing in sensitive user contexts or brand-unsafe environments. It is a platform-side guarantee rather than an advertiser-side control.
Can I exclude topics or keywords from my ChatGPT ads? No. There is no advertiser-side exclusion list. Context hints guide matching toward relevant conversations, but they are not exclusions and do not guarantee delivery in specific conversations.
Where will my ChatGPT ad not appear? No ads run beside conversations about personal health, mental health or politics, or in what OpenAI calls sensitive user journeys. Ads are also not shown to users on Plus, Pro or any Business plan, or to accounts identified as belonging to under-18s.
Can I see which conversations my ad appeared in? No. Advertisers receive aggregated performance information. Conversations are private and there is no adjacency reporting.
Does ChatGPT ads support third-party brand safety verification? No integration with third-party verification vendors is documented.
Does ChatGPT use my broader chat history to target ads? Ad selection considers the current conversation and, when ads personalisation is enabled, select signals from the user's broader ChatGPT experience.
Sources
| Claim | Source | Tier |
|---|---|---|
| Policy is to place ads only near chats that are safe, appropriate and aligned with user trust and brand safety | OpenAI Help Center, Ads in ChatGPT: The Basics, updated ~8 Aug 2026 | Confirmed, primary |
| Safeguards designed to prevent appearance in sensitive user contexts or brand-unsafe environments | Same | Confirmed, primary |
| Selection considers current conversation and, with personalisation enabled, broader ChatGPT signals | Same | Confirmed, primary |
| Ads appear below ChatGPT responses | Same | Confirmed, primary |
| Ads not shown to Plus, Pro or Business plan users, or accounts identified as under 18 | Same | Confirmed, primary |
| No ads beside personal health, mental health or political conversations, or sensitive user journeys | OpenAI ad policies v1.3, 15 July 2026 | Confirmed, primary |
| Advertisers receive aggregated performance information, not conversation access | OpenAI documentation and user communication | Confirmed and reported |
| Advertiser-side exclusion controls, adjacency reporting, third-party verification | Not documented as available | Absent |
Related reading
- ChatGPT ads context hints — what guides matching, and what it does not do
- Can my industry advertise on ChatGPT? — the category tiers and the sensitive-context layer
- Who should not buy ChatGPT ads yet — brand-safety sensitivity as a disqualifier
Changelog
15 August 2026. v1.0 published.
Field kit
Tools
Site
OpenAI (primary)
The InPromptAds Team
Writes about AI-native advertising and the shift from search to conversation.
