ChatGPT has started including advertising — so what might this mean for fertility treatment?

OpenAI began testing ads in ChatGPT in the US in February 2026 and has now expanded ChatGPT Ads into 31 European countries.

OpenAI says conversations are kept private from advertisers, user data is not sold, adverts are clearly labelled and kept separate from ChatGPT’s answers, and advertising does not influence the answers ChatGPT provides.

For advertisers, the system works in several familiar ways. Businesses can pay based on how many times an advert is shown or when someone clicks on it. They can also optimise campaigns towards outcomes such as purchases, registrations or other business goals. Advertisers can use geo-targeting to reach people in particular locations and custom audiences to reach groups of users using the targeting tools available on the platform.

OpenAI says tens of thousands of marketers have already advertised on ChatGPT. Ads are not shown to everyone. They are currently limited to people using the Free and Go plans. Plus, Pro and Enterprise subscriptions remain ad-free.

Fertility treatment advertising

What is potentially different about advertising in ChatGPT is the amount of context contained within a conversation. A fertility patient using a traditional search engine such as Google might search:

“Best IVF clinic London”

or

“IVF success rates over 40.”

In ChatGPT, the same patient might instead type:

“I’m 39 and have had two unsuccessful IVF cycles. My AMH is low, I’m trying to decide whether to stay with my current clinic or move elsewhere, I want to understand whether PGT-A is worth considering, cost matters, and I would be willing to travel if there were a good reason.”

That gives much more context than a few search terms. It includes the patient’s age, treatment history, concerns, priorities, budget considerations and willingness to travel.

The important difference is that conversational AI can potentially understand the wider decision the person is trying to make, rather than simply responding to a few individual search terms.

However, under OpenAI’s current advertising rules, this type of personal fertility-treatment conversation should not be eligible to show ads. OpenAI classifies personal health conversations as sensitive user contexts. Its current policy says ads should not appear alongside personal health, mental health or other sensitive user journeys.

So if a patient types:

“I’m 39 and have had two unsuccessful IVF cycles. My AMH is low, and I’m trying to decide whether to stay with my current clinic or move elsewhere…”

ChatGPT can still answer the question, but an advert should not be placed beneath the response simply because the system has recognised that the person is considering IVF treatment. This is different from the way ChatGPT Ads can work in non-sensitive conversations.

OpenAI says its advertising system can consider the context and intent of the current conversation. Rather than relying only on an exact keyword, advertisers can provide what OpenAI calls “context hints” describing the types of conversations, topics or keywords in which their product may be relevant. The system can then select from eligible ads based partly on how relevant they are to what the user appears to be trying to accomplish.

For example, a conversation about buying a laptop might say:

“I need a lightweight laptop for travelling, mainly for writing and video calls, under £1,000, with good battery life.”

The advertising system could recognise that the user is actively considering buying a laptop. An eligible laptop advertiser could then have an advert displayed below ChatGPT’s answer. The advert would be labelled as sponsored and visually separated from the answer, and importantly, the advertiser does not receive the conversation itself. OpenAI says advertisers do not receive users’ chats, chat histories, memories, names, email addresses, precise locations or sensitive information. Advertisers instead receive advertising-performance information such as views and clicks.

What about Europe?

There are additional restrictions in Europe because health information is treated differently from ordinary consumer information.

In the EU, health data are classed as special-category personal data under GDPR. Article 9 starts from the position that processing this type of data is prohibited unless a specific legal exception applies.

The EU Digital Services Act adds another restriction. Online platforms are prohibited from presenting advertisements based on profiling that uses special-category personal data, including health information.

The UK has a similar data-protection framework through UK GDPR, the Data Protection Act 2018 and PECR. The Information Commissioner’s Office states that health information is special-category data and that additional legal conditions apply when it is processed. Online advertising involving tracking and profiling must also meet transparency and consent requirements.

OpenAI has added its own platform restrictions on top of this regulatory environment. Its current advertising policy states that health-service advertising outside the United States is generally prohibited. Healthcare advertising is currently being introduced gradually and subject to approval, with the health-service categories permitted at present primarily applying to approved US advertisers.

So there are currently two separate controls.

First: who is allowed to advertise? - Health-service advertising is currently highly restricted, particularly outside the US.

Second: where can an advert appear? - Even where an advertiser is eligible to advertise, ads should not appear alongside sensitive personal-health conversations.

Where does the future of fertility advertising lie?

OpenAI is not the first company to introduce advertising into generative-AI experiences. Microsoft already serves ads within Copilot, Perplexity has experimented with sponsored questions and paid placements, and Google is introducing conversational advertising into AI Mode in Search. Anthropic, by contrast, has publicly committed to keeping Claude ad-free.

At the moment, fertility patients in the UK and Europe have a degree of protection because health information is treated as sensitive personal data, OpenAI currently restricts health-service advertising outside the US, and personal-health conversations are excluded from ad placement.

But this area is developing quickly. Generative AI platforms are only beginning to work out how advertising will fit into conversational systems, and their commercial models, ad formats and policies are likely to continue evolving. The key question is what happens as these systems become more deeply involved in helping people explore options, compare services and make decisions.

There is also another issue: how users themselves may adapt to the rules. Social-media platforms are already full of people sharing prompts, “hacks” and different ways of phrasing questions to get different results from generative AI.

So how long will it be before TikTok, Instagram and other platforms are full of influencers sharing prompts designed to reframe sensitive questions, avoid triggering restrictions, or encourage users to phrase requests in ways that produce more commercially useful answers?

That does not mean those prompts will necessarily work. There is also an important distinction between restrictions on what a model will answer and restrictions on where advertising is allowed to appear. But it does raise a wider governance question. If platforms build protections around certain types of language or conversational context, how robust will those protections be if users are being actively taught to phrase questions differently?

If generative AI becomes a routine part of how patients choose clinics, treatments, tests or other fertility services, the distinction between providing information and creating commercial influence will become increasingly important.

For now, the protections are relatively clear. What is much less clear is how those protections will evolve as conversational advertising becomes a more established part of the next stage of generative AI.