ChatGPT is introducing advertising in 31 European markets. People using a free account or the lower-cost Go plan will see ads, while Plus, Pro, and Enterprise are set to remain ad-free. The change also matters to users in Switzerland, where a media report says the rollout is imminent.
Who will see ads and where they will appear
According to OpenAI, ads will be limited to the Free and Go plans. The inclusion of Go is the notable part: paying for ChatGPT will no longer automatically mean an ad-free experience. Plus, Pro, and Enterprise subscriptions are expected to continue without advertising.
The European expansion of ChatGPT Ads covers 31 markets. OpenAI lists countries including Germany, France, Italy, Austria, and the Netherlands. Switzerland is absent from the company’s examples, but Netzwoche cites an SRF report saying ads will start there on August 24, 2026. This leaves a small discrepancy between the official announcement and reporting about the Swiss launch.
Ads are supposed to be clearly labeled and kept separate from the chatbot’s responses. OpenAI also says advertising will not influence the answers ChatGPT generates. The supplied information does not specify the exact screen position, how often ads will appear, or whether their presentation will vary by device.
The setting is attractive to advertisers because people often explain their intentions in more detail than they would in a standard web search. OpenAI’s examples include planning a trip, choosing business software, and furnishing a home. Someone looking for family-friendly accommodation within a particular budget, for instance, could see an ad when their purchasing decision is already becoming fairly specific.
What OpenAI promises about data
In its announcement about the European advertising rollout, OpenAI says conversations will remain private from advertisers and customer data will never be sold. It also says ads will be separate from answers and users will have control over ad personalization. These are claims from the provider and have not been independently verified in the supplied sources.
At the same time, OpenAI is expanding the way advertisers can measure campaign results. Campaigns can be optimized not only for views or clicks but also for conversions, meaning desired actions such as purchases or registrations. Netzwoche names the OpenAI Pixel, an application programming interface for conversion data, and links to outside measurement services among the available tools.
This does not automatically mean that advertisers can read your conversations. It does mean that an advertising system must process information about ad delivery and effectiveness. The provided texts do not fully explain which individual signals the Free and Go plans use for personalization, regional targeting, or campaign measurement.
The most immediate price of free access is therefore advertising and your attention. There is also a less visible exchange: you are using a system that can target ads to audiences and measure outcomes. The sources provide no exact subscription prices for the ad-free alternatives, so the financial trade-off cannot be calculated from this material.
Why sensitive content remains a separate risk
Advertising is only one part of the privacy issue. A report about a German privacy gateway for AI prompts notes that several apparently harmless details may be enough to infer someone’s identity. The filter described in the report is intended to remove sensitive information before it reaches ChatGPT.
Common workplace tasks can expose such details quickly. If you ask the chatbot to draft an email and include names, contact details, or information about an internal dispute, you may be processing data about colleagues, customers, or relatives. The same problem arises when analyzing large documents containing contracts, employee information, or unpublished business figures.
In a parallel announcement, OpenAI presented stronger protections for eligible customers using its application programming interface (API), which lets other services access its models. Under Zero Data Retention (ZDR), meaning processing without subsequent storage, prompts and model responses are not supposed to be retained after a request or made available to OpenAI staff. Enterprise customer data is also not used for training unless the customer explicitly opts in.
A preview called Private Safety Processing is intended to detect patterns of misuse across multiple interactions without exposing the underlying content to OpenAI personnel. According to OpenAI’s description of ZDR for advanced models, content either remains on infrastructure controlled by the customer or is encrypted with customer-controlled keys. However, this offer concerns eligible API customers and should not be mistaken for a blanket privacy guarantee covering personal chats on the free plan.
TechCrunch frames the move as a contrast with Anthropic. According to its report, Anthropic’s policy permits sessions and conversations involving certain covered models to be retained for 30 days so that possible safety issues can be investigated. OpenAI’s approach is meant to identify longer-term misuse patterns automatically, but the report describes it as a preview for selected customers rather than a generally available safeguard for everyone.
Pros and Cons of advertising in ChatGPT
Pros:
- Lower barrier to entry – According to OpenAI, advertising can help fund free and lower-cost access to ChatGPT.
- Ad-free options – Plus, Pro, and Enterprise are expected to remain free of advertising.
- Visible separation – OpenAI promises clearly labeled ads outside the chatbot’s actual answers.
- Personalization controls – Users are supposed to have control over whether advertising is personalized.
Cons:
- Ads on a paid plan – The lower-cost Go subscription will not be ad-free.
- Unclear data details – The sources do not fully explain which signals are used to deliver and measure individual ads.
- Commercial context – Advertising appears while you are comparing products or preparing a decision.
- Trust gap – The main promises about privacy and separation come from the provider and are not independently verified here.
This trade-off arrives amid broader skepticism about Artificial Intelligence (AI). A Pew survey summarized by TechCrunch found that 52% of respondents in the United States were more concerned than excited about the growing use of AI in daily life, compared with 37% in 2021. Another survey cited in the same report found that more than 70% of Americans thought AI was advancing too quickly.
Those figures cannot be transferred directly to Switzerland. They do help explain why an ad inside a chatbot may feel more sensitive than a conventional banner on a website. When a service processes personal goals, workplace problems, or purchasing criteria, acceptance depends heavily on whether the line between assistance, data use, and advertising remains credible.
What this means in practice
If you are a beginner, your first step is to check which plan you use and what ad personalization setting is available. Do not treat Free or Go chats like a confidential vault: remove names, email addresses, customer numbers, internal project names, and other details that identify a person or organization. For help drafting a message, an anonymous description such as “Customer A is complaining about a delayed delivery” is usually enough.
Do not share your ChatGPT login merely to split the cost of an ad-free subscription. With account sharing for ChatGPT, Claude, and Gemini, other people receive your login credentials and may also be able to view visible chat history. The report also identifies unclear responsibility, possible policy violations, and less accurate personalization of answers as risks.
If you are an advanced user, separate personal, professional, and organizational use cases. Companies should define which data may be entered into outside AI services and consider technical filters or anonymized templates. ZDR and Private Safety Processing may be relevant to eligible API customers, but they do not replace internal policies or a close review of the specific service being purchased.
Swiss organizations also need to consider privacy, trade secrets, and information concerning third parties. The sources do not provide a definitive legal assessment for Switzerland. In practice, the central questions remain whether a service is approved for a particular type of content and whether employees know which details must be removed before submitting a prompt.
Advertising does not automatically make ChatGPT unsafe, but it changes the exchange behind free access: the system now monetizes your attention as well as providing a tool for your inputs. The promised separation of ads and answers and the continued availability of ad-free plans are positives, while the exact data signals behind personalization and measurement remain unclear. The largest risk is still less likely to be the visible ad than sensitive information unnecessarily pasted into a chat or exposed through a shared account.
Sources
- OpenAI bringt Werbung in ChatGPT nach Europa – Netzwoche, 2026-08-19
- ChatGPT Ads expands across Europe – OpenAI, 2026-08-18
- OpenAI seeks to one-up Anthropic with new customer privacy protections – TechCrunch, 2026-08-19
- Offering Zero Data Retention for frontier models – OpenAI, 2026-08-19
- Privacy Gateway: Dieser Filter entfernt sensible Daten, bevor sie ChatGPT erreichen – t3n, 2026-08-19
- Account-Sharing bei ChatGPT, Claude und Gemini: 6 Gründe, warum du deinen Chatbot-Zugang nicht teilen solltest – t3n, 2026-08-18
- AI was supposed to win people over by now — it hasn’t – TechCrunch, 2026-08-19
