Leak Suggests that Ads Are Coming to ChatGPT Conversations. Users Have Feelings About It

For nearly three years, ChatGPT has been that rare thing on the modern internet: wildly popular and completely ad-free. That honeymoon looks close to over. New code buried inside the latest beta of ChatGPT’s Android app strongly suggests OpenAI is preparing to inject advertising into the chatbot’s interface. Starting, most likely, with search-style and shopping queries.

The clues come from version 1.2025.329 of the Android beta. Computer engineer Tibor Blaho dug into the app bundle and found strings like “ads feature,” “AdTarget,” “search ad,” “search ads carousel,” and “bazaar content.” Together, they look less like an experiment and more like the scaffolding of a full ad-serving system: targeting hooks, sponsored content wrappers, and carousel placements. None of this is visible to regular users yet, but it’s active in the test build.

OpenAI hasn’t formally announced an ad product for ChatGPT, but there have already been glimpses of what the future might feel like. In one widely shared screenshot, startup founder Yuchen Jin appeared to see a promotional card surface in the middle of a conversation. In another incident, a Peloton app suggestion popped up for a paying ChatGPT Pro subscriber, sparking backlash from users who assumed ads had quietly arrived. OpenAI later insisted the Peloton mention wasn’t an ad but a clumsy “app suggestion” test. A distinction that may not matter much to users once real sponsors are involved.

If ads are indeed coming, the timing is not mysterious. OpenAI is under heavy financial pressure: reports suggest the company faces no clear path to profitability through 2030, with compute and infrastructure demands running into the hundreds of billions of dollars. At the same time, ChatGPT’s weekly user base has swelled to around 800 million, up from roughly 300 million a year ago. A gigantic free audience that costs real money to serve. Subscriptions help, but they are unlikely to cover that bill alone.

Interestingly, this pivot comes after CEO Sam Altman repeatedly described ad-supported AI as “uniquely unsettling” and a “last resort,” warning that a system paid to recommend worse options like “a worse hotel above a better hotel” would be catastrophic for user trust. More recently, though, he has sounded more open, even praising some Instagram ads as “kind of cool” and saying he’s “not totally against” advertising inside ChatGPT if it’s done right.

So what might “ads done right” actually look like? The leaked strings, search ads, ads carousel, bazaar content, point toward card-style, clearly separated commercial units that appear alongside answers, especially when users ask about products, travel, or other high-intent topics. In the best case scenario, it won’t be like a pop-up banner, but more like a Google Search ad or shopping module embedded inside an AI reply. Early reports suggest the first wave would focus on free-tier users and on search-like use cases, not on every casual chat.

For users, the impact cuts both ways. On the positive side (we hope), a sustainable ad model could keep a powerful free version of ChatGPT alive, rather than pushing everyone toward paywalls or watered-down models. Highly relevant, clearly labeled ads tied to a query (“compare ultrabooks under $1,000”) might even feel useful, especially if they save time and come with strict guardrails around data use.

On the negative side, the risks are obvious and serious. Once a conversational assistant can be paid to recommend things, every answer about products, services, or even news becomes suspect. The line between “this is my best, neutral answer” and “this is sponsored” must be bright, loud, and non-negotiable. Users will also want explicit controls: the ability to opt out of personalised targeting, to turn off certain formats, and for paying subscribers, a guarantee of an ad-free experience.

For now, all we have is code in a beta build, a few awkward “not-actually-ads” app suggestions, and a lot of speculation. But in an industry where Google, Microsoft and others are already weaving ads into AI search and chat, it would be more surprising if ChatGPT stayed ad-free forever. The real test won’t be whether ads arrive. It will be whether OpenAI can introduce them without breaking the fragile trust that made people talk to a chatbot in the first place.

Nishant Kumar