ChatGPT will let you lock individual conversations behind a PIN or a biometric scan, such as a fingerprint or face check, and pull them out of the places where chats normally appear.
The feature, called Locked Chats, does two things. It adds a gate: opening a protected conversation will require the PIN or biometric. It also adds a hiding step: locked conversations will be removed from your regular chat history, your pinned chats and conversation search.
The second part is the more useful one. A passcode matters little if a conversation's title still sits in the sidebar or surfaces when someone types a keyword into search. Removing it from all three means a glance at your screen, or a borrowed device, turns up nothing to find.
The approach is familiar. WhatsApp's Chat Lock moves chosen conversations into a separate folder behind a fingerprint, face scan or passcode. ChatGPT's version follows the same logic for a product where people often type personal questions and drafts they would rather keep to themselves.
As described, though, this governs who can see a conversation, not what happens to it. It controls access and visibility on the screen in front of you. That makes it a privacy tool for shared devices and curious people nearby, which is a real problem, but a narrower one than the word "protect" suggests.