OpenAI has launched a new feature in the “ChatGPT” application called “Locked Chats”, which gives users the ability to hide their private conversations behind a secret code, protecting them from the eyes of others when using the same phone or device.
According to the company, the locked conversation is hidden from the usual chat history, pinned conversations, and search results, and can only be opened with a four-digit code, or through biometric authentication such as “Face ID” on iPhone and Android phones.
To activate the feature, the user opens the options menu next to the desired conversation, selects “Lock the conversation,” then confirms the decision, before creating a PIN for the account or entering it if he has created it previously. The conversation then moves to the “Locked chats” section in the “Projects” list.
The company explained that the conversation may appear temporarily in open tabs or on other devices until the lists are updated, and its content may remain displayed in the current tab on the web after it is locked. Therefore, users are advised to update applications and check other devices before sharing the screen or lending the device.
OpenAI stressed that the feature does not provide end-to-end encryption, nor does it exempt it from its legal obligations related to data. You also don’t automatically change settings for using conversations in model training or data retention policies. Locked conversations remain subject to your account settings and company policies.
The company warned that resetting the PIN leads to the deletion of all locked conversations and files uploaded to this space, and the deletion cannot be undone, and the support team cannot recover the content.
The feature is being rolled out gradually, and it may take several days for it to appear in accounts, including free accounts and subscribers to the “Go,” “Plus,” and “Pro” plans. The feature works with voice mode on iPhone and Android, but it does not allow you to lock Codex conversations in the same way.