ChatGPT privacy: data controls, temporary chat, and what is shared

ChatGPT has a switch that decides whether your conversations help train OpenAI's models. This page explains it from OpenAI's own help page.

What we read

OpenAI's Help Center page "Data controls in ChatGPT", read in a browser on 14 November 2026 (the page said it was updated 3 days earlier). Settings and wording change; check the page itself.

The main switch

The page says that on personal workspaces with ChatGPT Free, Plus or Pro, data sharing is enabled by default. The switch is called Improve the model for everyone. To turn it off on the web: open your account menu, choose Settings, then Data controls, turn off Improve the model for everyone and choose Done. On iOS and Android, open the sidebar, tap your profile to open Settings, choose Data controls and turn it off. When you are signed in, the setting applies across web and mobile. When you use ChatGPT without signing in, the same switch is in Settings and the choice is saved in that browser.

The page adds that voice mode has a separate control, and that Codex has its own "Include environments" setting.

What turning it off does not do

Turning off the switch does not delete or hide saved chats. Regular chats stay in your history, archived chats stay archived, and project chats stay in their projects. To remove content, delete chats or the account separately.

Temporary chat

A temporary chat, while it remains temporary, does not appear in your history, does not create or update memories, is not used to improve OpenAI models, and may be retained for up to 30 days for safety purposes. If you save a temporary chat, it becomes a regular chat and follows your account's settings.

Feedback is an exception

The FAQ says: if you give feedback such as thumbs up or thumbs down on a response after opting out of training, the entire conversation associated with that feedback may be used to train OpenAI models. So do not click the feedback buttons on conversations that contain anything private.

Memory is a different setting

The page keeps memory and model training separate: memory controls personalisation of responses; training controls whether content can improve models. Turning off memory does not turn off safety features that may use limited, safety-relevant context in rare, high-risk situations.

Export and delete

Signed-in Free, Go, Plus and Pro users can request a copy of their data from ChatGPT settings or the privacy portal, and can delete the account from settings. Deleting an account is permanent and affects its data and subscriptions. These options are not self-service in managed workspaces; ask the workspace owner.

Business, Enterprise, Edu and the API

By default OpenAI does not use content from ChatGPT Business, Enterprise, Edu or Healthcare workspaces to train its models, and does not use inputs and outputs through the API Platform for training by default. Managed workspaces can have their own retention and access settings, and your organisation's policies still apply.

A safe setup for personal use

1. Turn off Improve the model for everyone if you do not want training use.
2. Use a temporary chat for anything sensitive, and still keep private identifiers out.
3. Avoid the feedback buttons on sensitive chats.
4. Delete old chats you do not need.

This is an independent guide. It is not connected with OpenAI, Anthropic, Google, Microsoft or any other AI company; product names belong to their owners. Statements about companies are taken from their own help pages as read on 13 and 14 November 2026. The prompts are our own wording and were not run against every tool. Not legal, financial or professional advice.