ChatGPT launches ads: a journalist asked 500 questions to see what the bot shows

OpenAI has done what many expected but were afraid to say out loud: ChatGPT now shows ads. Commercial ads have appeared on the free tier in the US, and a Wired journalist decided not to wait until they reached him. He ran his own experiment: 500 questions to the bot, 500 chances to see how the ad machine works inside the world's most popular AI chatbot.

OpenAI confirmed the ad launch in early 2025. The company had long hinted that subscriptions are not the only path to monetization. Generative AI is expensive to run: the billions of tokens processed every day require enormous computing resources. ChatGPT's free users, who number in the hundreds of millions, have until now used the service without any direct monetization from advertisers. That has now changed.

The journalist methodically asked ChatGPT 500 questions across a wide range of areas: finance, health, travel, technology and everyday household tasks. The goal was to find the pattern: how closely the ads relate to the context of the query, how intrusive they are, and how they are built into the interface. This is not a superficial test, since 500 questions give a statistically significant sample.

The results were revealing. Ads do not appear every time, and they do not appear randomly: they are tied to the topic of the conversation. Travel questions triggered ads from travel services, and financial queries triggered ads for banks and investment platforms. The ads are also neatly designed. They do not interrupt the answer; they appear next to it and do not imitate its content. OpenAI appears to be trying to balance commercial pressure against user trust.

One of the main concerns about launching ads in AI assistants is that the bot will start to "recommend" paid products inside its answers, disguising advertising as advice. So far that is not happening: the separation between ads and the answer holds.

Still, many questions remain. How exactly does OpenAI pass query data to advertisers? Is data stored for targeting? The company has provided minimal information about the technical side of the system. That creates a zone of uncertainty: users do not know whether their query becomes a signal for a real-time ad auction, as happens in Google's search engine.

The potential audience is enormous. ChatGPT is one of the fastest-growing products in the history of the internet. Hundreds of millions of people entrust it with their questions, tasks and ideas every month. Turning that traffic into ad inventory is a logical business step, but it has consequences for how users perceive AI answers. Even if ads do not seep into the text, their mere presence changes the context.

OpenAI's competitors are taking a wait-and-see position for now. Google Gemini is built into the ecosystem of a company that earns trillions from advertising. Anthropic, Meta* and other players have not yet announced plans to launch ads in their AI assistants. But if OpenAI's experiment proves financially successful, the industry will most likely follow the same path.

ChatGPT's ad monetization is a signal of where the generative AI industry is heading. Free access to a smart AI assistant no longer implies that the user is the customer. As in the search-engine era, they once again risk becoming the product.

*Meta has been designated an extremist organization and is banned in Russia.