OpenAI ChatGPT Ads Advertising Transparency and Advertiser Disclosures 2026

Transparency in ChatGPT ads runs in three directions. What OpenAI tells users about an ad. What it tells advertisers about their results. And what it tells everyone else about who is advertising. The first two are well documented. The third is mostly missing, and that gap is why brands end up building their own record.

Everything below comes from OpenAI's own pages, read on 2026-10-21, plus one outside review by the Network Advertising Initiative. Where a page did not say something, we say so instead of guessing.

What users are told

OpenAI's user FAQ (Ads in ChatGPT) makes these commitments:

Ads appear below a ChatGPT response, are clearly labeled, and are shown separately from the answer. OpenAI is also testing a multi-image ad format during image generation.
Ads run on systems separate from the chat model. Advertisers cannot shape, rank, or alter answers.
An ad is a paid placement, not an endorsement.
Each ad has a three-dot menu with Hide ad, Report ad, About this ad (why you are seeing it), and Ask ChatGPT (which shares that one ad with the chat, since ChatGPT cannot see your ads by default).
Settings, then Ad Controls, lets eligible users turn off personalization, stop past chats and memory from informing ads, view ad history and topics, and delete ad data.

Who sees ads is also disclosed. Ads may appear on the Free and Go plans. Plus, Pro, Business, Enterprise, and Edu do not get ads, and accounts identified as under 18 do not see them. Temporary Chats show none. Ads are not eligible near sensitive topics such as personal health, mental health, or politics.

One caveat from an outside reviewer. The Network Advertising Initiative reviewed ChatGPT ad placements with screenshots from May 2026 (NAI report). It found the units marked as sponsored and set apart from the answer, but noted that a "Learn about ads and personalization" link appeared under some placements and not others. Its point: a disclosure a user only sometimes sees does less to build understanding.

What advertisers are told

For advertisers, OpenAI publishes the mechanics (The Basics):

Format. Advertiser name, favicon, title, copy, landing page, and an image.
Selection. Expected relevance and outcomes, using the context and intent of the conversation, the ad's landing page and copy, advertiser-provided context hints, and personalization signals when enabled.
Auction. Relevance-weighted, second-price. Advertisers set a maximum CPM or CPC bid per ad group. OpenAI suggests starting CPC bids at $3 to $5.
Reporting. Impressions, clicks, spend, CTR, average CPC, average CPM, and conversions.

And it publishes the limits on what advertisers get back. Advertisers never receive chats, chat history, memories, name, email, precise location, IP address, or sensitive information. They receive aggregated performance data such as views and clicks.

What is disclosed about advertisers

This is the thin direction. OpenAI's ad policies describe how advertisers are vetted:

Advertisers must give accurate information about business identity, ownership, and affiliations, and may not impersonate another brand.
Reviews happen at three levels: the advertiser at sign-up, the creative and landing page at upload, and the placement before serving. Most review is automated with human oversight. Restricted categories such as financial services can require enhanced verification or manual review.
Ad accounts require business verification (API Organization Verification).

The policy page also keeps a changelog. Version 1.1 (April 2026) stopped blocking medical, legal, and financial advice contexts by default. Version 1.3 (July 2026) added the advertiser policy section and clarified eligible financial and health categories and markets.

What we did not find on any of these pages: a public, searchable record of which advertisers run which ads in ChatGPT. Google has the Ads Transparency Center, where you can look up an advertiser's ads by region, date, and format. Meta has its Ad Library. OpenAI's help center and policy pages describe the About this ad panel for a user looking at one ad. They do not describe a library of all ads.

The checklist for brands

Given those three directions, here is what a brand team should actually do.

Read the policy before assuming silence. If your category is restricted, an empty ad slot on your prompts can reflect policy, not demand.
Do not rely on About this ad for competitive research. It explains why one user saw one ad. It is not a search tool.
Do not expect Ads Manager to show rivals. It reports on your campaigns.
Build your own ad record. Run your buyer prompts on a schedule, capture every sponsored unit, and keep the history. That is the advertiser disclosure OpenAI does not publish, scoped to the questions that matter to you.
Keep paid and organic separate. OpenAI promises ads do not change answers. Your reporting should reflect that: an ad under an answer is a different event from a citation inside it.

The tool that fills the gap

Promptwatch Ads Radar is the closest thing to an advertiser disclosure record for your own prompts. It captures sponsored ads inside ChatGPT Search, Google AI Overviews, and Google AI Mode answers, with the advertiser name and domain, title, description, image or price, target URL, and the prompt that triggered it (Ads Radar docs). Open any captured response and the Ads and Citations tabs list paid and organic sources separately, which is exactly the separation OpenAI's own principles call for.

The reason to pick Promptwatch over an ads-only feed is what sits next to the ad. The same prompts carry your mentions, sentiment, and page-level citations, and crawler logs tell you whether ChatGPT's bots fetched your page at all. Ads Radar is on the Business plan ($579/mo) and every agency plan (from $199/mo for Kick-off), per llms.txt on 2026-10-21. Essential and Professional do not include it.

Transparency in ChatGPT ads is real at the level of the single ad. At the level of the market, you have to watch it yourself.