OpenAI rolled out a new chat-locking feature on Tuesday, letting users secure sensitive conversations behind a four-digit PIN, Face ID, or Touch ID. The tool moves private chats out of regular history, though it stops short of end-to-end encryption or wiping out past memory.
When OpenAI announced the update on Tuesday, it brought a feature long familiar to apps like WhatsApp into the AI assistant space. The tool is designed for moments when users need to hand their phones to family or friends, or simply want to keep private discussions away from wandering eyes on crowded public transit. Instead of lingering in the main feed, secured threads shift to a separate folder.
Users Hide Conversations with a PIN or Biometrics
To tuck a conversation away, users tap the three-dot menu inside an active chat. Once secured, the application relocates the thread away from standard history, pinned conversations, and search results. On desktop and mobile, these hidden projects rest neatly under the sidebar menu.
Titles remain invisible until unlocked. Access requires a four-digit account PIN, while mobile users can rely on facial recognition or fingerprint scans.
Plan Eligibility and Gradual Rollout Across Accounts
The privacy update is reaching a wide array of tiers simultaneously. Accounts on Free, Go, Plus, Pro, Business, Enterprise, and Edu plans are eligible for the feature. However, healthcare deployments miss out on the upgrade, and temporary chats cannot be locked.
Because the deployment is happening in stages, the option might not appear immediately for every user. The company noted in support documentation that locked chats are rolling out gradually. If you don’t see Lock chat in an eligible chat’s more options menu (•••), the feature may not be available for your account yet.
Memory Isolation and Security Caveats to Keep in Mind
While the tool keeps casual observers out, it does not rewrite how the underlying software handles data storage. OpenAI explicitly notes that locking a conversation does not create an encrypted vault, alter retention schedules, or remove information that ChatGPT remembered before you locked the chat.
Workspace compliance obligations and safety requirements remain fully active. If a user has opted into model training, conversations inside the folder can still be accessed for that purpose.
On the memory front, locked threads maintain their own isolated recall network. New messages written inside a secured project update that specific shared memory without bleeding details into regular chats.
With Memory enabled, your locked chats project has its own memory, shared between the chats inside it. New messages you send after locking a chat can update that memory, but do not update your regular Memory, so that new information from your locked chats doesn’t leak into regular chats.
OpenAI, via PCMag Australia
Users must also manage their PINs carefully. Forgetting the code requires a full reset, which results in the permanent deletion of every locked chat, file, and protected document tied to that previous PIN.
