Is it true?
Is it true that ChatGPT now puts ads in its answers?
By the end, you'll know what OpenAI has announced about advertising in ChatGPT, who it affects, what other assistants like Copilot, Gemini and Meta AI do, and how to tell an ad from an answer.
Translated with AI from the Spanish original. Nobody has reviewed this translation; if anything seems off, the Spanish version is the reference.
Partly. Free ChatGPT and Go show ads, in Spain too since August 2026. According to OpenAI, they sit at the bottom of the answer and are labelled. Whether they influence what it says has not been verified.
Where the claim comes from
The claim has been going around since January 2026. That month, OpenAI, the company that makes ChatGPT, announced it would start testing ads in its assistant. The test began in the United States on 9 February 2026, with adults using the free version or ChatGPT Go, a paid plan that is simpler than the others. In August 2026, OpenAI expanded it to more countries, including 31 European markets, Spain among them.
The strongest version of the claim says something more worrying: that ChatGPT slips advertising into what it tells you. In other words, if you ask it for a recipe or shopping advice, the answer could be tilted towards a brand that has paid. That concern isn't absurd. It's how many free internet services make money, and an assistant that talks with you knows a lot about what you're interested in.
What the evidence says
What is documented is that OpenAI has decided to show ads in ChatGPT. It's not a rumour. The company itself explained it in January 2026, in a post about how it planned to introduce advertising.
According to OpenAI, ads appear below the answer, separate from it and labelled as sponsored. They aren't mixed into the text. The company also says that ads don't influence what ChatGPT answers, that it doesn't sell your conversations to advertisers, and that it doesn't show ads to people it identifies or estimates to be minors. It also says ads don't appear near conversations about health, mental health or politics.
Be careful with those promises. They are what OpenAI says about itself, not something that can be checked from the outside. No one outside the company has been able to review how the system decides which ad to show or whether that affects the answers.
What OpenAI did make clear is who isn't affected: according to the company, the higher paid plans (Plus, Pro and the business plans) don't carry ads. Advertising stays in the free version and in Go.
And in other assistants
ChatGPT isn't the only assistant with advertising. Microsoft has shown ads in some answers from its chat-based search engine since 2023. That service is now Copilot. There, sponsored links appear next to the answer when the question is about something you can buy.
Google has confirmed ads in its AI Mode, the version of its search engine that answers in conversation. On 28 October 2026, it announced it was bringing them to ten more countries, including Australia, India and Singapore. Spain isn't on that list. In the United States, it also shows them in follow-up questions, the ones you ask after the first answer. According to Ars Technica, Google acknowledges that AI search earns it less money than traditional search.
Two things need to be kept apart: Google's AI-powered search engine and the Gemini app. What Google has confirmed is advertising in search. Maunga hasn't found an official Google announcement saying the Gemini app shows ads.
Meta, the owner of Facebook, Instagram and WhatsApp, took a different route. It doesn't put ads inside Meta AI. Since December 2025, it has used what you talk about with its assistant to choose which ads it shows you in its other apps. According to The Register and other outlets, in October 2025 Meta initially left the European Union, the United Kingdom and South Korea out of this change.
Google has a similar case. Since 16 October 2026, its privacy policy says content from Gmail, Drive or Google Photos can be used to personalise ads. That only happens if you choose to let other Google services use that data. Before, the policy said plainly that it didn't do this.
What we don't know
The most important thing we don't know is whether ads influence the answers. OpenAI says they don't, but there is no independent audit to confirm it. An audit would be a review carried out by someone outside the company with access to how the system works.
It's also unclear which countries see ads in each assistant today. Companies switch them on gradually and region by region, and what was true a few months ago may have changed. What is confirmed is that, since the end of August 2026, OpenAI has shown ads in ChatGPT in 31 European countries, Spain included, to people using the free version or Go.
Finally, screenshots of supposed ads in ChatGPT are circulating, for example blocks with headings like «Esto también podría interesarte» ("You might also be interested in this"). Maunga hasn't been able to verify that they are official ads rather than normal suggestions from the assistant or doctored images. Something recommending a brand isn't enough to tell whether it's advertising.
How to tell if what you're seeing is an ad
Look for the label. Ads in Copilot, in Google search and, according to OpenAI, in ChatGPT are marked as "sponsored" or "ad". They also appear separate from the text of the answer. If a recommendation doesn't carry that label, it isn't a paid ad, though that doesn't mean it's neutral: assistants also make mistakes or repeat whatever appears most on the internet.
If you don't want ads in ChatGPT, OpenAI offers the higher paid plans, which according to the company don't carry them. With Meta, according to the press, the only sure way to stop what you say to Meta AI from being used to choose ads is not to use the assistant, except where the law gives you another option. With Google, as well as checking the personalisation settings in Mi Centro de Anuncios, disconnect Gmail, Drive and Fotos from Personal Intelligence, the feature that connects your Google apps to its assistant.
Watch out: An ad being kept separate from the answer doesn't guarantee the answer is neutral. When you ask an assistant to recommend a product or service, compare it with other sources before you decide.