ChatGPT Privacy Settings: What Each Control Does
Training, history, memory and connected apps are separate ChatGPT controls. Review each one before sharing personal or customer information.
BMR Digital · Source review 2026-11-06 · UK English
Start with the account you actually use
ChatGPT privacy settings do not form a single private mode. The first useful step is to establish whether you are using a personal account, a work workspace or a school-managed account. The rules can differ. Open the account menu and check the signed-in identity rather than assuming that a familiar browser means a familiar account. A shared computer may have another person’s conversation history or connected services. On a personal account, review the settings yourself; on a managed account, ask the responsible administrator about the applicable data terms. Write down the account type and date of your check. This gives you a useful starting point if a setting changes later. Avoid copying workplace information into a personal account while waiting for clarification. The convenience of a saved login is not evidence that your employer has approved the service.
Model training is one decision, not every decision
OpenAI describes the Improve the model for everyone control in Settings and Data controls. Turning it off stops new eligible conversations being used to improve models under that control. It does not mean the service stops processing your messages to answer them, nor does it instantly remove everything you previously shared. Review OpenAI’s current explanation for exceptions and feedback-related handling. If you want to discuss a problem containing identifying details, changing the training setting is not a substitute for removing those details first. A useful test is to ask whether the question still works with invented names, approximate dates and a simplified example. Often it does. Save your own explanation of why you selected a setting, not a screenshot containing your email or conversation titles. Recheck after a major product update or when switching accounts.
Chat history and deletion have their own role
A conversation appearing in your sidebar is about account history and convenient access. Do not confuse that with permission to use its content for training. Deleting a visible chat is also different from preventing future use of new chats. Check the provider’s current deletion and retention wording before making assumptions about how quickly content disappears from all systems. For everyday housekeeping, remove conversations you no longer need and avoid leaving a sensitive thread open on a shared screen. Keep any records you genuinely need in an appropriate protected location, not solely in a chatbot sidebar. If a conversation contains a leaked password, deleting the chat does not restore the password’s secrecy. Change the credential and consider where else it may have been copied. The response should match the information exposed, rather than relying on the delete button as a universal remedy.
Memory can be separate from a conversation
Memory and personalisation features may remember useful facts across conversations. Review the current memory controls and the stored items where your account provides that view. Turning a memory feature off should not be assumed to delete existing stored memories. Likewise, removing a chat should not be assumed to remove a fact saved elsewhere. Work through both controls and check the official instructions for deleting remembered information. A reasonable household rule is not to ask an assistant to remember children’s full names, addresses, medical details or a regular travel routine. Even a helpful fact can become an unnecessary disclosure when combined with others. Before sharing a device with someone else, consider what answers could reveal from the account’s saved context. Memory is a convenience setting, not a secure personal filing cabinet or a guarantee that every future answer is confidential.
Temporary Chat has limits worth understanding
OpenAI says Temporary Chat does not appear in chat history and is not used for model improvement. Its guidance also describes limited safety retention, including that a copy may be kept for up to 30 days. That means temporary is not the same as never stored anywhere. Read the current Temporary Chat page before using it for an unusual situation. The safer habit remains to avoid sharing unnecessary personal data in the first place. A temporary conversation can be useful when you do not want a routine topic added to your visible history, but it should not encourage uploading a bank statement, identity document or confidential client file. If a tool, connector or external action is involved, consider the receiving service’s terms as well. You cannot decide how another organisation treats data simply by choosing a mode in the chatbot interface.
Connected apps need a separate permissions check
Connected services can give an assistant access to email, documents or other account material. Review each connection rather than leaving every integration enabled because it was useful once. Check what the connection can read or do, which account is connected and whether its purpose still applies. Disconnect services that you no longer need, and review permissions in the original service too. A document folder may contain considerably more information than the single file you intended to discuss. For work tasks, get approval before connecting business email or shared drives. For personal tasks, avoid connecting a household member’s account by accident on a shared device. Do not assume that disconnecting reverses past data transfers. Where an external service has received information, its own retention and deletion processes may matter. Limiting access before a task is usually easier than trying to undo a broad permission afterwards.
A practical monthly review
Set aside a short review that you can actually repeat. Confirm your account identity, check training preferences, inspect memory, remove obsolete chats and review connections. Then check login security, including the strongest authentication options available for the account. Keep your device updated and avoid sharing account passwords. If someone sends a supposed privacy-fix link, navigate to the service yourself rather than signing in through an unexpected message. Record that the review happened and note any controls you could not find. Labels and availability change, so a missing setting is a reason to consult the current help page, not to assume a screenshot from an old article still applies. For a printable routine spanning several providers, compare the AI Safety & Privacy Playbook. The free checklist offers a smaller starting point. Neither resource replaces the provider’s current terms or your organisation’s approval process.
Official sources and further checks
Provider settings can change and differ by account, plan and location. Check the current official instructions for the service you actually use. This is independent educational guidance, not a safety or legal compliance guarantee.
A printable next step
Compare The AI Safety & Privacy Playbook for the stated printable guidance, or start with the free checklist. Only genuine uploaded files are offered for download.
View The AI Safety & Privacy PlaybookUseful AI safety reading
All resourcesChatGPT Privacy Settings: What Each Control Does
Training, history, memory and connected apps are separate ChatGPT controls. Review each one before sharing personal or customer information.
AI Safety for Children: A Parent's Starting Guide
Begin with age rules, private information, supervised chatbot use and calm family conversations about deepfakes and misleading answers.
AI Policy for Small Business: Simple UK Starter Guide
Set approved tools, prohibited data, human checks and incident reporting before staff use generative AI for customer work.
AI Voice-Cloning Scams: How to Check a Fake Call
Do not rely on a familiar voice alone. End the call and independently contact the person or business through a known route.
What Not to Share With AI: Practical Examples
Keep passwords, identity documents and unnecessary personal or confidential details out of AI prompts. Use fictional examples and minimise data.
Gemini Activity & Privacy Settings Explained
Review Gemini Keep Activity, deletion, connected services and account type separately; switching off activity does not mean no processing or retention.
Claude & Copilot Privacy Controls Compared
Compare account type, training preferences, history, memory and connections rather than assuming Claude or Copilot is completely private.