ChatGPT has been quietly building a profile on you. Not just what you asked it, but what it figured out about you that you never said out loud.

Every conversation, it takes notes. Your habits. Your job. Your worries. How you think. Most people have no idea it is even there, let alone how much detail it has pieced together.

Here is how to see it for yourself, and how to take control of it in under two minutes.

Step 1: Ask for everything it has stored

Open a brand new chat and paste this in exactly:

Raw list prompt — copy this exactly

Lay out every single thing you've stored about me as one complete list. Don't summarize anything.

Most people have no idea this list exists. Read it slowly. Don't be surprised if there are things on it you had forgotten you ever mentioned.

Step 2: Ask what it figured out on its own

Stay in the same chat and send this:

What-tipped-you-off prompt — copy this exactly

Now tell me everything you figured out about me that I never told you, and what tipped you off.

This is the one that surprises most people. The second half matters most: when ChatGPT has to say what tipped it off, you can see exactly which of your own words led to each guess.

Step 3: Go deeper with the original reveal prompt

Open ChatGPT and paste this in exactly:

Reveal prompt — copy this exactly

Based on everything you know about me, tell me what you've saved and what you've inferred, including things I never told you directly.

Read the answer slowly. This is not a guess — it is pulling from the actual memory ChatGPT has stored about you across every conversation you have had.

Step 4: Push it one step further

Once you have the first answer, send this follow-up:

Follow-up prompt — copy this exactly

What assumptions have you made about my income, my personality, and my goals? Be honest, even if it's uncomfortable.

This is the part that catches most people off guard. ChatGPT will often name assumptions about your financial situation, your personality traits, and what it thinks you are working toward, based purely on patterns in how you have talked to it.

The first time you run this, it can feel a little unsettling how accurate it is.

Step 5: See the actual list

Everything ChatGPT just described to you lives in one place. To see the raw list:

  • Tap your name in the bottom left, then go to Settings
  • Tap Personalization
  • Tap Memory, then Manage memories

This is the full list of everything ChatGPT is holding onto about you — every saved memory, in plain text.

Step 6: Clean it out

From that same Memory screen, you have three options:

Delete a single memory. Tap the memory you want gone and remove just that one entry. Useful if most of what it knows is fine, but one or two things feel too specific.

Wipe everything and start clean. Clear the entire memory list in one action if you would rather ChatGPT start from zero.

Turn Memory off entirely. Right in the same menu, toggle Memory off so it stops saving anything new going forward.

One more setting worth checking

While you are in there, look under Data Controls. There is a separate toggle called "Improve the model for everyone," which controls whether your conversations can be used to train future versions of ChatGPT. Turn that off too if you would rather your chats stay out of that pool entirely.

Your ChatGPT privacy checklist

Run through these five once, and you are done:

  • Read what it has saved. Settings → Personalization → Memory → Manage memories.
  • Delete anything you do not want it to know. One memory at a time, or all of them.
  • Decide about Memory itself. Leave it on if it helps you, turn it off if you would rather it start fresh every time.
  • Turn off "Improve the model for everyone" under Data Controls.
  • Check back every few months. Memory keeps growing, so run the Step 1 prompt again once in a while.

Why this matters

None of this is necessarily malicious — Memory exists to make ChatGPT more useful to you over time. But most people never asked for a running file to be kept on their habits, income, and personality, and most people never checked what is actually in it.

Your AI should work for you, not keep notes on you that you have never read. Now you know exactly where to look, and exactly how to clear it out whenever you want.