OpenAI updates privacy policy for ad personalization on ChatGPT

OpenAI has updated its privacy policy to account for ad personalization as it expands sponsored content within ChatGPT to new regions worldwide.

How the ad rollout unfolded

OpenAI first began testing ads in the US on February 9, 2026. The initial experiment targeted adult users on the Free and Go subscription tiers, keeping Pro, Business, and Enterprise plans ad-free.

By August 11, 2026, the program had expanded to the UK, Mexico, Brazil, Japan, and South Korea. The US privacy policy received ad-related data handling updates around April 30, 2026. A separate policy update on June 2, 2026, addressed requirements specific to the EU and UK.

In the US, ads are served by default, with users given the option to opt out. In the EU and UK, the opposite applies: personalized ads require explicit user consent before they can be shown.

What the policy actually says

The updated privacy framework clarifies several categories of data that can be used to serve relevant ads. These include chat context, ad interactions, and, for eligible users, stored memories within ChatGPT.

Users can toggle ad personalization on or off at any time through their settings and can delete ad-related data whenever they choose. Ads within ChatGPT are clearly labeled, and the company states they do not influence the AI’s actual responses.

Advertisers will not have direct access to user conversations. OpenAI handles the targeting internally, using contextual signals to match ads to users without exposing the raw dialogue.

OpenAI communicated these changes through emails sent to its user base, walking users through the new controls and explaining their significance.

Privacy tensions and what to watch

In the US, the opt-out model means most users will receive personalized ads unless they actively dig into their settings and flip a toggle. In the EU and UK, the opt-in requirement means OpenAI will likely see lower personalization rates among European users.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.