Dear OpenAI Team,
I would like to propose an important privacy and security feature for ChatGPT Memory.
ChatGPT can remember personal information from previous conversations, including private experiences, family details, and sensitive matters. While this feature is extremely useful, users should have stronger control over when this information can be revealed.
My proposal:
- Introduce an optional identity verification feature before ChatGPT reveals sensitive information from past conversations.
- Allow users to create a private security question, PIN, or use biometric authentication where supported.
- Let users decide which categories of remembered information require additional verification.
- Prevent the disclosure of protected memories until the user successfully verifies their identity.
- Provide an easy way to enable, disable, and manage these privacy protections in ChatGPT settings.
Why this matters:
Someone with temporary access to an unlocked device or an active ChatGPT session could ask questions about the account owner’s private history.
An optional verification system would give users greater confidence that their personal memories remain protected.
This feature should be implemented securely, without relying on ChatGPT itself to validate a secret answer through ordinary conversation.
I hope OpenAI will consider adding this feature to improve user privacy, security, and trust.
Thank you for listening to user feedback.