🔒

3 prompts for basic privacy with AI

Copy-paste into ChatGPT or Claude, in order, in the same chat.

Prompt 1

Using only this conversation and any saved memory actually available to you, list the personal information you know about me.

Group it into: identity, location, work, finances, health, relationships, and other sensitive details.

For each item, show what I shared, where you found it, and whether it is an explicit fact or an inference. Flag details that could identify me when combined.

Do not guess or invent anything. Do not claim to inspect chats or account records you cannot access.

Give me the report directly, not instructions for creating one.

Prompt 2

Rewrite my most recent request into a version that reveals less about me while still getting a useful answer. Replace names, companies, addresses, contact details and account numbers with placeholders. Generalize dates, locations and amounts. Remove secrets entirely. Check combinations of details that could still identify me. Return the ready-to-copy prompt first, then list what you changed.

Prompt 3

Create a standalone version of the useful conversation that I can copy and share. Keep the useful question, answer and conclusions. Remove personal disclosures, identifying details, confidential material and sensitive information about other people. Check both my messages and your replies. Return only the cleaned-up version, with clear placeholders where necessary.

Bonus: turn off training

I use [ChatGPT / Claude] on [web / iPhone / Android / desktop], with a [plan name] account. Using current official documentation, show me how to turn off the use of my conversations for model training. Give me the exact setting to check and explain any feedback-related exceptions. Don’t claim you changed anything for me. Wait for me to confirm before continuing.