A map of settings, not a new master setting

ChatGPT now has a Privacy Center. The name might suggest one button that makes an account private. It is not that. OpenAI describes the new hub as a place to understand privacy options and find the existing settings that manage them. Opening it changes nothing on its own.

The rollout covers signed-in Free, Go, Plus, Pro and Business users on the web and on iOS and Android. It does not include Enterprise, Edu or ChatGPT for Healthcare. OpenAI says access may still depend on an account, app version or region, so a missing menu item need not mean a user has lost their other controls.

The choices do different jobs

A useful first stop is model improvement. Eligible personal accounts can turn off 'Improve the model for everyone' in Data controls. OpenAI says that stops new conversations from being used to improve models, but it does not remove those conversations from chat history. Business and other managed workspaces have different defaults and administrator rules.

Memory is separate. It concerns information ChatGPT may retain or reference to personalize future replies. Its settings sit under Personalization. Deleting a chat does not automatically erase a saved memory from it. A user who wants to remove both needs to check both places.

Temporary chats and connected apps

The hub also points to Temporary chat. OpenAI says these conversations stay out of chat history, do not create or update memories and are not used to improve models while they remain temporary. They are not a promise of instant erasure: the company says it may retain a copy for up to 30 days for safety. A personalized temporary chat can still use existing memories and other context unless personalization is turned off when the chat starts.

Connected apps are another control. The access an app has depends on what the user granted at connection time and on any workspace restrictions. Changing when ChatGPT asks permission is not the same as revoking an app's access; disconnecting it is a separate action. App permissions also do not replace memory, retention or model-improvement settings.

What to check in practice

On the web, OpenAI says to open the account menu, then Help, then Privacy center. On mobile it appears in Settings when available. The hub groups topics under personalized chats and ads, chat privacy, and data use and access. Its Manage links take a user to the relevant existing setting.

For a quick audit, ask four separate questions: Is model improvement on? What memories are saved? Which apps remain connected and what can they do? Do I need a temporary chat for this conversation? The answers depend on plan and workspace. The new hub makes those questions easier to find. It cannot answer them for the user or override an organization's policies.

Sources

  1. OpenAI Help Center: Privacy Center in ChatGPTCurrent primary product documentation for rollout, account eligibility, navigation, topics and limits of the hub.
  2. OpenAI Help Center: How OpenAI handles data in consumer servicesCurrent primary documentation distinguishing model improvement, memory, temporary chat, retention and managed-workspace defaults.
  3. OpenAI Help Center: Connected apps in ChatGPTPrimary documentation distinguishing app permissions from provider access, disconnection and general privacy controls.