ChatGPT Ads Reach Europe on 24 August — and GDPR Made Them Ask First

Ten days ago we noted that when OpenAI switched on ads inside ChatGPT, one region was conspicuously missing from the list: Europe. That gap has now closed. From 24 August, ads arrive for free-tier users across thirty-one European markets, including the EU and Switzerland. But the more interesting story is not that Europe caved — it is how differently the ads had to be built to be allowed in at all.

What Changes on 24 August

On Tuesday 18 August, OpenAI confirmed that ChatGPT advertising will expand to thirty-one European countries — the European Economic Area plus Switzerland — starting 24 August. It is the largest single regional expansion since the US pilot began in February, and it brings ads to the big European markets that had been held back: Germany, France, Spain, Italy, and the rest. As everywhere else, ads appear only on the free tier and the low-cost Go plan (eight euros a month); Plus, Pro, Business, Enterprise, and Education remain ad-free. The commercial logic is unchanged and unhidden: OpenAI is burning cash ahead of an IPO that could value it near a trillion dollars, and nearly a billion weekly users on free tiers are the obvious place to find revenue.

The GDPR Difference: Ask, Don't Assume

Here is what makes the European launch genuinely different, and it is worth understanding precisely. Under the GDPR, personalised advertising counts as processing personal data, which means OpenAI needs a lawful basis to do it. Many ad platforms lean on the vaguest available justification — so-called legitimate interest — to target users without asking first. OpenAI has chosen the harder path: at launch in Europe, ads are not personalised at all. Targeting uses only the conversation topic, approximate location, device type, time of day, and language — context, not profile. Personalisation will come only later, and only for users who actively opt in through advertising controls in their account settings. Advertisers receive only aggregate figures. OpenAI Ireland is the data controller, which puts the Irish Data Protection Commission in the supervisory seat. This is, almost line for line, a more restrained product than the one Americans received — and the reason is not OpenAI's generosity. It is European law.

Better, or Just Politer?

That difference is real, and it should be credited. But honesty requires asking whether consent-based advertising is genuine control or merely a legal ritual. The uncomfortable truth about consent is that most people click through it. A prompt asking permission to personalise ads, buried in settings, is easy to accept and easy to ignore, and platforms have spent a decade perfecting the art of making yes the path of least resistance. Contextual-only targeting at launch is meaningfully more private than profiling — but the phrase doing the heavy lifting is at launch. The architecture for personalisation is being built; it simply waits behind a checkbox. And beneath all of it sits the same structural fact we named ten days ago: the most intimate consumer technology ever made now has a commercial sponsor, and the wall between what the assistant knows and what the advertiser can eventually reach is only as strong as the consent flow protecting it. Europe has made OpenAI ask first. It has not made the underlying trade disappear.

Claim and Counter-Claim

The generous reading is that this is the Brussels Effect working exactly as intended. Faced with the GDPR, OpenAI did not just delay the European rollout; it redesigned the product — consent-first, contextual-only, opt-in for anything more. European users are getting a materially more privacy-respecting version of ChatGPT ads than anyone else on earth, and they are getting it because European rules forced the question. That is regulation shaping a product for the better, in real time, for hundreds of millions of people.

The skeptical reading is that consent-first is the most sophisticated compliance strategy available, not necessarily the most protective outcome. Building on explicit consent from the start actually positions OpenAI to serve more targeted ads later, legally and durably, once users have clicked yes — a firmer foundation than the shaky legitimate-interest approach that got Meta into trouble. Consent can be a genuine safeguard or a liability shield with a friendly face, and which one it is depends entirely on how real the choice feels at the moment it is offered. Add the unresolved question we have raised before — that the publisher content feeding ChatGPT's answers remains unpaid while OpenAI sells the ad space beside it — and the picture is of a monetisation machine being assembled carefully, within the rules, rather than restrained by them. The honest synthesis: the European version is better, the improvement is real, and it is also the beginning of the model, not a limit on it.