OpenAI expanded ChatGPT Ads to the United Kingdom, Mexico, Brazil, Japan, and South Korea on August 11, moving its advertising pilot into five additional markets as the company tries to build a revenue stream around conversations where users are comparing options, researching products, and making decisions.
The launch is not a small product-label change. It makes ChatGPT a more serious advertising surface outside the United States, Canada, Australia, and New Zealand, and it gives brands a new way to reach users inside an AI assistant rather than only through search results, social feeds, retail media, or display networks.
OpenAI is still framing the program as a controlled test. In its announcement, the company said ChatGPT’s answers remain separate from ads, conversations are kept private from advertisers, and users retain controls over the ad experience. The harder test begins as the pilot becomes a larger commercial channel: OpenAI must prove that sponsored placements can be relevant without making users question whether the assistant is nudging them toward a paying advertiser.
Who sees ChatGPT ads now
The current advertising experience is aimed at logged-in adult users on lower-cost or free ChatGPT plans. OpenAI’s advertiser help page says ads are not shown to Plus, Pro, or Business-plan users, and are not shown to accounts identified as belonging to users under 18. The company also says ads appear below ChatGPT responses rather than inside the model’s answer.
That placement matters. A search engine can put sponsored links above organic links and still operate inside a familiar advertising model. A chatbot has a different trust problem because the interface feels conversational and advisory. Users may ask for help choosing a laptop, planning travel, comparing insurance, finding software, or deciding what to buy for a business. In those moments, the boundary between recommendation and advertisement has to stay visible.
OpenAI says each ad includes an advertiser name, favicon, headline, description, landing page, and image asset. Its help page also says ad delivery considers the current conversation, the ad’s landing page, headline, copy, advertiser-provided context hints, and, when ad personalization is enabled, selected signals from the user’s broader ChatGPT experience. Those hints are not exact-match keywords and do not guarantee placement in specific conversations.
What advertisers are buying
OpenAI’s ads pitch is built around intent. The company’s ChatGPT Ads site tells businesses they can reach people while they explore options, compare alternatives, weigh tradeoffs, and make decisions. That is the same commercial territory that made search advertising valuable, but with a more personal interface and a richer stream of conversational context.
The buying mechanics are also starting to look more like a conventional ad platform. ChatGPT Ads supports reach campaigns priced on CPM and click campaigns priced on CPC. OpenAI recommends that CPC advertisers start with a maximum bid of $3 to $5 per click, and says the system uses a relevance-weighted second-price auction. Ads Manager Beta reporting includes impressions, clicks, spend, click-through rate, average CPC, average CPM, and conversions.
Those details move ChatGPT Ads beyond a brand experiment. They give performance marketers the basic pieces they expect from a measurable channel: objectives, bids, creative, reporting, conversion tracking, and persistent URL parameters for analytics. The early question for advertisers is not only whether ChatGPT has enough reach, but whether a click from an AI conversation behaves more like a search click, a social click, or something new.
Why trust is the product risk
OpenAI says advertisers do not receive a user’s chats, chat history, memories, or personal details. It says advertisers receive aggregate performance information such as views and clicks, and that ads are not eligible to appear near sensitive or regulated topics including health, mental health, or politics during the test. Users can dismiss ads, give feedback, see why an ad was shown, delete ad data, and manage personalization.
Those controls are necessary because conversational advertising raises a different kind of scrutiny. A recent academic audit of early ChatGPT ads collected more than 3,000 ads from simulated U.S. accounts and found that lower-income accounts were more likely to receive ads, regardless of race, while the ads observed in the first phase were clearly separated from response text and mostly skewed toward consumer goods. The study’s authors cautioned that ad systems in large language models need ongoing outside monitoring as formats and targeting mature.
That finding does not mean OpenAI’s ad system is broken. It does show why the company cannot rely only on design promises. If ads are selected using conversational intent, personalization signals, and auction dynamics, the public record will need more than a visible sponsored label. Audits, advertiser transparency, targeting limits, and easy user controls will determine whether the ad business feels like a helpful layer or a quiet pressure inside a trusted assistant.
The bigger platform shift
The timing also connects to a wider shift in how AI assistants are becoming places where users act, not just places where they ask questions. ChatGPT can already help with shopping-style research, app-connected workflows, and, as of this week, restaurant reservations through OpenTable, Resy, and Yelp in supported markets. Ads fit into the same direction: AI assistants are becoming commercial interfaces.
For OpenAI, advertising can subsidize free and low-cost access while helping pay for expensive model serving. For users, it changes the mental model of ChatGPT. A free assistant that occasionally shows labeled ads is familiar enough. A deeply personalized assistant that helps choose products, plan purchases, or book services while an auction runs in the background is more sensitive.
The August 11 expansion does not settle that tension. It widens the test. ChatGPT Ads now has more countries, more advertisers, and more real-world usage from which OpenAI can learn. The strongest version of the product would make sponsorship obvious, keep organic answers independent, give users meaningful control, and let researchers verify whether delivery is fair. Anything less would make monetization feel like the thing users came to ChatGPT to escape.
