The short answer
- See it: Settings › Personalization › Memory, then read the memory summary. Or ask ChatGPT "What do you remember about me?"
- Edit it: type a correction into the summary, or highlight a passage and correct it.
- Delete it: remove the item and the chat it came from. Deleting the chat alone isn't enough.
- Turn it off: ••• › Delete and turn off memory. Your chats stay.
- Skip it once: start a Temporary Chat and pick Unpersonalized.
Every step below comes from OpenAI's help page Memory in ChatGPT, read on 25 October 2026. OpenAI notes that controls vary by plan, region, platform and workspace, so if a setting isn't on your screen it may not be available for your account yet.
You saw "Memory updated"
That note means ChatGPT saved or changed something it will use to personalize later answers. To find out what, open the memory settings below. When personal context shapes an answer, a Sources list can also appear under the reply: from there you can open the past chat, saved memory, file or email it drew on, and correct or delete a saved memory directly.
How to see what ChatGPT remembers, step by step
1. Open the memory settings
Go to Settings › Personalization › Memory. Depending on your plan you'll see some of: Memory, Reference saved memories, Reference chat history, Memory summary, Manage and Saved memories. On a linked teen account, a parent or guardian manages these through parental controls; in a work or school workspace, an admin may have turned some of them off.
2. Read the memory summary
The memory summary is ChatGPT's high-level view of what it may use about you. It draws on past chats, saved memories, custom instructions, files in your Library and connected apps such as Gmail, depending on plan and region. It is not a complete list: OpenAI says it can leave out details that are less relevant or come from sources it doesn't show one by one. Asking ChatGPT "What do you remember about me?" is the quickest way to check the rest.
3. Correct what's wrong
Type the correction into the summary's text field, or highlight the passage and give the correction. You can also just tell ChatGPT in a chat that something is outdated. Where it's offered, Don't mention this again makes ChatGPT stop raising a detail, but it doesn't delete the chat, file or email the detail came from.
4. Delete a memory so it really goes
Where a memory lives decides how you remove it. To remove something completely, delete it from each place it appears:
- the memory summary or the saved memories list,
- the chat where you shared it, including archived chats,
- files in your Library,
- connected apps that hold it (disconnect the app to stop future access).
Deleting a chat does not delete a saved memory created from it, and deleting a saved memory doesn't remove mentions of it from old chats. Changes can take time to spread, and OpenAI may keep logs of deleted saved memories for up to 30 days.
5. Turn memory off, fully or partly
- Everything: open the ••• menu on the memory summary and choose Delete and turn off memory. It deletes what the summary holds and switches Memory off. Past chats stay, and if you switch Memory back on, ChatGPT may build new memory from them.
- Just old conversations: turn off Reference chat history. What ChatGPT learned from past chats is scheduled for deletion within 30 days; your saved memories stay.
- Saved memories too: turning off Reference saved memories also turns off Reference chat history.
6. Skip memory for one conversation
Start a Temporary Chat. It never creates or updates memory. Before it starts, choose Unpersonalized if you also don't want it to use your existing memory and custom instructions; you can't change that once the chat begins. If you save a temporary chat, it becomes a regular one and follows your normal settings.
Saved memories vs the memory summary
ChatGPT has had two memory designs. Saved memories is the older one: a list of individual facts you can scroll, edit and delete one at a time. Improved memory keeps a continually updated summary instead. Where your plan offers both, switch in Settings › Personalization › Memory: choose Saved memories for the list, or Try improved memory to go back. Switching doesn't delete your chats.
Privacy: what else to check
Anything sensitive you tell ChatGPT can end up in memory, so review the summary and the Sources under answers now and then. On personal accounts, OpenAI may use chats and remembered information to improve its models while Improve the model for everyone is on; turn it off in Data controls. Business, Enterprise, Edu and Healthcare workspaces aren't used for training by default.
When ChatGPT's memory isn't enough
ChatGPT's memory stays inside ChatGPT: OpenAI documents no export, and Claude, Cursor or Gemini can't read it. If you switch between assistants, keep what matters in a store outside all of them. BrainTube is one: you save videos, articles and notes once, and ChatGPT, Claude and Cursor read the same library over MCP. Setup is in how to give your AI long-term memory with MCP.
More on how ChatGPT's memory works and why it says "memory is full": does ChatGPT remember your conversations? How it compares with Claude and Gemini: which AI has the best memory?