On February 9, 2026, OpenAI began testing ads in ChatGPT in the United States. The scope was tightly drawn: logged-in adult users on the Free and Go tiers only; Plus, Pro, Business, Enterprise, and Education plans see no ads. By August 11, the test had gone live in the UK, Mexico, Brazil, Japan, and South Korea, after Canada, Australia, and New Zealand joined in March. From one-country test to eight markets in six months — this is no longer an experiment; it is a product line.
What the Ads Look Like and How They’re Chosen
Format-wise, ads are labeled “sponsored,” visually separated from conversational answers, and OpenAI states plainly that ads “do not influence” responses. Selection runs on three signals in parallel: the current conversation topic, past chats, and the user’s history of ad interactions.
The privacy lines are drawn more strictly than on most platforms: advertisers get aggregate-level reporting only, with no individual identities; minors (whether self-reported or predicted) are fully excluded; and no ads appear near sensitive topics like health, mental health, or politics.
The Cost of Opting Out
Free-tier users can turn off personalized ads — but OpenAI is explicit about the price: opting out reduces your daily free message allowance. Watching ads is, directly, how you pay for the free tier. The controls are otherwise complete: report, give feedback, see “why am I seeing this ad,” delete ad interaction data in one tap, and review ads history.
Results and Expansion Rhythm
The March 26 update disclosed early numbers: ads had “no impact” on consumer trust metrics, with dismissal rates staying low — the basis OpenAI cited for international expansion. May brought the announced plan for the UK and four more markets; August 11 confirmed all five live, with more markets promised within the year. On the business side, openai.com/advertisers opened its doors. The long-term direction is stated too: ad formats, objectives, and buying models will keep evolving.
The Attention Economics of a Chat Interface
For product people, three things from these six months are worth remembering. First, keyword matching becomes “conversational intent matching” in a chat interface — intent signals are stronger than keywords, which is both the real selling point of AI-era ad products and the core of the privacy debate. Second, the “pay to remove ads” tiering has stabilized: free and low-cost tiers see ads, premium tiers stay clean — a structure many products will copy. Third, trust metrics are the ceiling of any ads product; OpenAI publicly citing “no impact on trust” signals it knows that the moment ads degrade answer quality, users will switch. Making the free tier ad-supported is, in essence, trading user trust for scale — a balance every team doing ad monetization should be watching.
Sources
AI-assisted summary compiled from the sources above, reviewed by a human before publishing.