How to Remove Your Personal Data from ChatGPT (2026 Guide)
Delete ChatGPT chats, use Temporary Chat, turn off model improvement, or request personal-data removal through the OpenAI Privacy Portal.
The correct ChatGPT privacy control depends on what you want to change: future use of new conversations for model improvement, a saved chat or file, a temporary conversation, or personal information appearing in ChatGPT responses. These controls are separate and none promises to erase information from every website, search engine, or AI system.
This guide reflects OpenAI's current Data Controls guidance, Temporary Chat FAQ, chat and file retention policy, and personal-data-removal process. Interfaces and eligibility can change, so confirm the live OpenAI instructions immediately before submitting a request.
For a general request workflow that is not specific to OpenAI, see how to request data removal. Keep the provider-specific OpenAI instructions and boundaries on this page as the source of truth for ChatGPT controls.
Quick answer
- Stop new consumer conversations being used to improve models: turn off Improve the model for everyone in ChatGPT Settings → Data Controls. OpenAI says this applies to the account, not just one device.
- Keep a new conversation out of history: use Temporary Chat. It does not appear in history and is not used to improve models, but OpenAI may keep a copy for up to 30 days for safety.
- Delete an existing conversation or file: delete the chat from ChatGPT, then check the separate Library or project/custom-GPT storage area for files that need their own deletion action.
- Stop personal information such as a phone number appearing in ChatGPT responses: use the official OpenAI Privacy Portal and select Remove my personal data from ChatGPT responses. OpenAI reviews these requests case by case; approval is not guaranteed.
Which ChatGPT privacy control should you use?
| What you want to change | Use this control | Important boundary |
|---|---|---|
| Future use of new consumer conversations for model improvement | Settings → Data Controls → Improve the model for everyone | This is forward-looking; it does not delete existing chats or remove an answer already generated elsewhere. |
| A new conversation that should not appear in history | Temporary Chat | It may still be retained for safety for up to 30 days; connected third-party actions have their own policies. |
| A saved conversation | Delete the chat in ChatGPT | OpenAI says deletion removes it from the account immediately and schedules permanent deletion within 30 days, subject to stated exceptions. |
| A saved file | Review the separate Library, project, or custom-GPT storage area | Deleting a chat does not necessarily delete a file saved separately. |
| Personal information in ChatGPT responses | OpenAI Privacy Portal → Remove my personal data from ChatGPT responses | This does not remove the source page, search-engine result, or another AI provider's output. |
Step 1: Turn off model-improvement use for new conversations
- Open ChatGPT on the account you use.
- Open Settings.
- Select Data Controls.
- Turn Improve the model for everyone off.
OpenAI says that new conversations will then not be used to improve its models, and that the setting applies to the account across devices. This is not a deletion request for previous chats, uploaded files, or personal information already appearing in a response.
Tired of dealing with data exposure?
Choose relevant provider workflows, review the generated drafts in your browser, and send or submit each request yourself. Matching, eligibility, and provider requirements still need checking.
Source: OpenAI Data Controls guidance.
Step 2: Use Temporary Chat for a new sensitive conversation
Start a new chat and select the Temporary option shown in ChatGPT. OpenAI says Temporary Chats do not appear in history and are not used to improve its models. Non-personalized Temporary Chats do not use memory and do not create new memories by default; a personalized Temporary Chat may use existing memories or custom instructions. OpenAI may still keep a copy for up to 30 days for safety.
If a GPT uses an action or sends information to a third party, that recipient's privacy policy may allow a different retention period or use. Temporary Chat is therefore a control for the ChatGPT conversation, not a universal privacy guarantee for connected services.
Source: OpenAI Temporary Chat FAQ.
Step 3: Delete chats and review stored files separately
OpenAI says chats remain in the account until you delete them. When you delete a chat or account, it is removed from the account immediately and scheduled for permanent deletion from OpenAI systems within 30 days, unless it has already been de-identified or must be retained longer for legal or security reasons.
Also check for separate copies:
- Delete the conversation from ChatGPT.
- Review the Library and delete files stored there when applicable.
- Review files attached to projects or custom GPTs, which may follow separate retention rules.
- Review Memory or personalization settings separately; deleting a chat should not be treated as proof that every saved memory has been removed.
Source: OpenAI chat and file retention policy.
Before you submit a personal-data request
Keep the request specific and share the minimum information needed to identify the issue:
- Separate the target: an account chat, uploaded file, saved memory, or personal information appearing in a response may use a different control.
- Record the exact prompt, response excerpt, date, and any conversation or source link that the current form requests. Do not include unrelated chats or another person's information.
- Check the source page separately. If the response repeats a wrong or outdated page, a correction or removal request to that source may be needed in addition to an OpenAI request.
- Use the official `privacy.openai.com` domain and follow its current verification instructions. Do not send an identity document to an address found in an unsolicited message or search result.
This evidence helps you describe the narrow issue without turning a privacy request into a broader disclosure of personal information.
Step 4: Request removal of personal data from ChatGPT responses
Use this route when ChatGPT produces personal information about you that is inaccurate, excessive, irrelevant, or no longer appropriate. It is different from deleting a conversation and different from turning off model improvement.
- Open the official OpenAI Privacy Portal.
- Select Make a Privacy Request.
- Choose Remove personal data from ChatGPT responses.
- Describe the specific information and where it appears. Include relevant chat links or URLs when the current form requests them.
- Explain why the information is inaccurate, excessive, irrelevant, or no longer appropriate.
- Follow the current verification instructions. OpenAI says a government-issued ID or alternative proof will be required; submit only through the official portal and only what the current process requests.
OpenAI assesses these requests case by case. Its published factors include accuracy, public or professional role, sensitivity, impact, public interest, the source of the information, and whether the information remains relevant. A request can be declined.
Source: OpenAI personal-data-removal guidance.
What an OpenAI request does not remove
An OpenAI response-removal request does not remove the source page where information was published, a Google or Bing result, a data-broker profile, or another AI provider's output. Those are separate targets with separate request processes.
If the source is a people-search or data-broker page, use the provider's current route in the OfflistMe directory, preserve the submission confirmation, and verify the listing separately. A ChatGPT request and a data-broker opt-out should never be treated as one combined deletion action.
Frequently asked questions
Does turning off model improvement delete old chats?
No. It changes the treatment of new conversations going forward. Delete existing chats separately, and use the Privacy Portal for personal information appearing in ChatGPT responses.
Is Temporary Chat completely private?
No. Temporary Chats do not appear in history and are not used to improve models, but OpenAI may retain a copy for up to 30 days for safety. Third-party actions connected to a GPT may have separate policies.
Does deleting a chat delete uploaded files?
Not necessarily. OpenAI says chats and Library files are managed separately. Review the Library, projects, or custom GPT storage area when a file also needs to be deleted.
Can OpenAI deny a personal-data-removal request?
Yes. OpenAI says these requests are reviewed case by case and may be declined when other legal, public-interest, accuracy, or contextual factors outweigh the request.
Does OpenAI remove my information from Google or other AI tools?
No. OpenAI states that response removal does not remove information from external websites or search engines. Contact each separate source or service when that is the actual target.
How do I remove my phone number from ChatGPT?
Use the official OpenAI Privacy Portal and select Remove my personal data from ChatGPT responses. Identify the phone number, the response or chat where it appears, and any relevant source or URL the current form requests. OpenAI reviews these requests case by case; this does not remove the source page, a search-engine result, or another AI provider's output.
What should I include in a ChatGPT personal-data request?
Describe the exact information, the response or conversation where it appears, and the reason for the request using only the details the current OpenAI form requires. Keep unrelated personal information, other people's data, and unnecessary identity documents out of the submission.
Privacy checklist
- [ ] Turn off Improve the model for everyone when you do not want new conversations used for model improvement.
- [ ] Use Temporary Chat for a new conversation that should not appear in history.
- [ ] Delete the chat and then check separate Library, project, custom-GPT, and Memory surfaces.
- ] Use the official [OpenAI Privacy Portal for response-level personal-data requests.
- [ ] Keep the request narrow and submit only the verification information the current official process requires.
Related guides
- How to Remove Personal Data from Google Gemini
- How to Opt Out of AI Training Data
- Facial Recognition and AI Search Privacy Guide
- How to Remove Your Name from Google
- Complete Data Broker Removal Guide
- OfflistMe Data Broker Directory
Sources and review boundary
This guide summarizes the linked OpenAI help pages as reviewed on August 31, 2026. OpenAI may change product labels, retention details, eligibility, or request fields. The linked first-party pages control over this summary.
Understand your privacy rights
Where a privacy right is relevant, these plain-English explainers show what each law covers and what to verify before making a request.
Related Data Broker Removal Guides
Take back your privacy today
Review provider-specific routes, prepare your requests locally, and send or submit each one yourself.
Review Provider RoutesFree to review provider routes · Optional one-time unlock from $9.00 · No subscription