Sam Altman confirmed that OpenAI is already starting to test ads in ChatGPT and that the experiment will affect both free users and those on the ChatGPT Go budget plan. The idea, according to the approach the company has shared regarding this test, is to allow more people to use AI without always having to pay, while OpenAI builds a sustainable business model over time.

Ads in ChatGPT: What changes for free and Go users?

The ad test will begin “in the coming weeks” and, at least in this first stage, will focus on adult users logged in from the United States. OpenAI has explained that ads will appear when there is a sponsorship “relevant” to the conversation and will be displayed at the bottom of the replies within the chat.

The point the company is emphasizing most is that advertising will not guide ChatGPT's responses: that is, the response should remain optimized to be helpful and not to sell you something. Furthermore, OpenAI also states that it will not share conversations or sell user data to advertisers, in a clear attempt to protect trust, which is literally the “oxygen” of a product like this.

“The most important thing is that we won't accept money to influence the response ChatGPT gives you, and we keep your conversations private from advertisers. We know that many people want to use a lot of AI and don't want to pay, so we hope that a business model like this can work,” Altman shared on his had previously expressed reservations about including ads in ChatGPT due to the risk of damaging the assistant's credibility, OpenAI is now moving towards a hybrid model where advertising helps sustain mass access. And it makes sense: maintaining a popular, fast, and increasingly capable AI costs a lot of money (infrastructure, GPUs, energy, research, security), so relying solely on subscriptions limits its reach.

In parallel, OpenAI also drew a line: the Plus, Pro, and Business/Enterprise plans will remain ad-free, reinforcing the idea of ??“paying to remove friction,” while the free or cheaper plans accept advertising in exchange. In practice, this move also acknowledges something the company can no longer ignore: most people want to use AI daily without paying (or at least without paying as much), and OpenAI needs a way to make that viable in the long term.

OpenAI insists that ads will be clearly marked and separated from the response logic, and that personalization will be optional (key to avoiding the feeling of constant surveillance). If the test works—and if the feedback doesn't become a storm—it could open the door to a new stage: more accessible assistants for millions, but with a permanent battle to maintain trust as the priority.