A New Level of Security: ChatGPT Adds PIN Protection for Private Chats

OpenAI is adding a feature called “locked chats” to give ChatGPT conversations extra protection. When you turn it on for a conversation, that chat disappears from your regular chat history, pinned chats, and search results. You can open it by entering a 4-digit PIN. On iOS and Android devices, you can also use face recognition, such as Face ID, to unlock chats. All hidden conversations are kept in a separate section called “Locked Chats.”

Purpose and privacy limits

This feature is mainly designed to protect private information from people who can see your screen or get physical access to your device. OpenAI points out several important technical and legal limits:

• The feature does not use end-to-end encryption.
• It does not hide information from law enforcement or employers when legal issues are involved.
• Locking chats does not automatically change your settings for AI model training or how long OpenAI keeps your data on its servers.
• If you reset your PIN, all locked chats and files uploaded to them are permanently deleted. OpenAI Support cannot restore this data.

Availability and rollout

OpenAI is releasing the new feature gradually, so it may take several days to appear in some accounts. It is available to free users and people with Go, Plus, and Pro subscriptions. On mobile devices, locked chats can also be used in voice mode. The only exception is conversations that use the Codex model, which cannot be hidden with this feature.