ChatGPT Locked Chats Add PIN Protection, but Privacy Gaps Remain
ChatGPT Locked Chats use PIN or biometrics, but TechRadar finds links, memory and open tabs can expose data.
To lock a conversation, users open the menu under the three dots in the upper right of the chat window, tap “Lock Chat” and confirm. If they have not already created a four-digit PIN, they must do so, or link the lock to the face or fingerprint authentication used to unlock the device. OpenAI creates a new Project specifically for locked chats, and the conversations disappear from regular history and search results unless unlocked, so someone browsing the account without authentication should not be able to find them.
Users must remember the PIN. They can change it later, but if they forget it and reset it, ChatGPT permanently deletes every locked chat and file protected by the old PIN, according to TechRadar.
TechRadar reported several caveats. A conversation can remain open in the same browser tab immediately after it is locked, and previously opened devices might temporarily display its title until they refresh. Users need to close or refresh the relevant tab and check other devices if they want the chat hidden. Unlocking a protected Project also gives temporary access, and the chat does not re-lock automatically unless the user selects Lock again after finishing.
Shared links also continue to work after a conversation is locked. If a user made a shared link to a ChatGPT interaction, locking the conversation afterward will not disable that link; the user must delete the shared link to close the loophole. Locking restricts access inside the account, but it does not recall copies that have already been shared.
Information already remembered by ChatGPT before a conversation is locked remains part of regular ChatGPT Memory and can influence later responses, TechRadar reported. The information is not automatically erased, and users must manually go into the memory section and remove it. After the lock is set, newly remembered details remain separate from ordinary ChatGPT memory and can be used within other locked conversations without leaking into other conversations. TechRadar advised users who expect to lock a chat to do so before sensitive discussions begin.
For especially sensitive discussions, TechRadar said Temporary Chat may be a more appropriate choice. Temporary conversations do not appear in regular history, do not create memories and are not saved by OpenAI for its model improvement program, though users cannot return to the conversation afterward.
Locked Chats are an appealing addition for everyday privacy concerns, according to TechRadar, but the feature is not comprehensive or entirely automatic. Controlling who can open a conversation is useful, but it does not address information embedded in ChatGPT’s memory or a shared link, and tracking that content takes more effort than relying on a PIN and facial recognition.
Editor's Summary
OpenAI’s Locked Chats add PIN or biometric protection to individual ChatGPT conversations, hiding them from regular history and search unless unlocked. TechRadar reported that shared links, pre-lock memory entries and open tabs can still expose information, while Temporary Chat offers a more limited but less persistent alternative.