ChatGPT ads reach the UK this year

OpenAI is bringing ads to ChatGPT. The company began testing them in the United States in February. On 7 May it named the UK in the next wave of the pilot.

Ads will appear for people on the free and Go tiers. Paid plans stay clear. Plus, Pro, Business, Enterprise and Education accounts will not show ads.

This lands at an awkward moment. AI has finally become reliable enough to trust. Now the question shifts to who is paying for that trust.

What is changing

The format is restrained, at least to start. Ads sit below the answer, not inside it. They are labelled as sponsored and kept visually separate from the response.

OpenAI says ads do not influence the answers you get. Ads run on separate systems from the chat model. Advertisers cannot rank, shape or alter what ChatGPT tells you.

The matching works on context. If you research recipes, you might see a meal kit or grocery delivery ad. The system uses your current chat, past chats and previous ad interactions.

Several guardrails are in place for the test:

You also get controls. You can dismiss an ad, ask why you saw it, delete your ad data and manage personalisation.

The trial has grown fast. OpenAI reports more than 600 advertisers and over £75m in annualised ad revenue within six weeks. Self-service ads are coming, which opens the door to far smaller advertisers.

The trust question

Consider the irony. AI spent two years earning trust. Models improved. Hallucinations dropped. People grew comfortable relying on AI for real decisions.

That comfort is the asset advertising now sits beside. The worry is no longer whether it will make something up. It is whether you are being nudged towards what someone paid for.

OpenAI is clear that ads and answers are separate. Early trial data backs this up. The company reports no measured impact on trust and fewer than 7% of ads rated as low relevance.

Take that as encouraging, not settled. The wall between a sponsored slot and an organic recommendation is easy to describe and hard to hold. The pressure to blur it grows with every revenue target.

My view: watch the boundary, not the promise. A labelled ad below an answer is honest. A product woven into the answer itself would not be. The first is fine. The second would break the thing that makes ChatGPT useful.

Why this matters for UK SMEs

This affects your team directly. Ads land on the free and Go tiers. That is exactly where most staff actually use ChatGPT.

Not everyone in a business has a paid seat. Plenty of people open the free app on a personal login. They draft emails, summarise documents and check facts on it daily.

Picture a sales rep researching a supplier on a free login. A sponsored result sits below the answer. It looks like guidance. It is a paid placement. The difference matters when a buying decision follows.

So the people seeing ads are often the same people you trained. The cost of getting AI wrong now carries a commercial layer that did not exist before.

The fix is not fear. It is knowing your setup. Three things are worth checking this week:

This is a governance question, not a technical one. It sits in the 80% that is people and culture, not the 20% that is tools.

A fair trade, done well

None of this makes ads wrong. Free AI has been heavily subsidised for years. Someone has to pay for the compute.

If ads keep a capable tool free for people who cannot pay, that is a reasonable deal. Access matters. A student or a sole trader getting ChatGPT for nothing is a good outcome.

The deal only holds on conditions. Ads must stay non-invasive. They must stay clearly separated. Users must keep real control. And they must stay off the paid plans people rely on for sensitive work.

Get it wrong and the cost is steep. OpenAI knows this. It shut down the Sora app in April after the AI video network failed to hold users.

Rivals are watching too. Anthropic ran adverts during the Super Bowl mocking the idea of ads inside AI answers. A clumsy rollout would hand that argument real weight.

What to do next

The lesson here is not about ads. It is about choice. The platform and tier you pick now shape what your team sees and trusts.

Model benchmarks get the headlines. For an SME, the practical decision is simpler. Which platform, which tier, which guardrails, and who owns that call.

gecco helps UK SMEs make those choices with a clear head. We are vendor-neutral across ChatGPT, Claude, Copilot and Gemini. We focus on the setup and the habits, not the hype.

If you want a clear read on where your team sits with AI adoption, take the AI readiness assessment.

Sources: OpenAI (Testing ads in ChatGPT; Ads in ChatGPT Help Center), Campaign, TechCrunch, Storyboard18.