OpenAI has introduced a new Locked Chats feature in ChatGPT, allowing users to protect sensitive conversations with an additional layer of privacy.

 

Announced on Tuesday, the feature lets users secure conversations using a four-digit PIN or biometric authentication, such as Face ID or Touch ID, on supported devices.

 

Unlike Temporary Chats, which don’t appear in regular chat history, Locked Chats remain accessible whenever users need them. The feature is designed for situations where people want to keep personal conversations private, such as when sharing their device with friends or family.

 

Once locked, a conversation moves to a dedicated Locked Chats section, separate from regular chat history, pinned conversations, and search results. This makes it easier to keep sensitive discussions apart from everyday interactions.

 

OpenAI has also introduced separate memory behavior for Locked Chats. When Memory is enabled, new information from conversations in this section can update its own shared memory without updating regular ChatGPT Memory.

 

However, locking a chat does not mean it is encrypted or completely hidden from every form of access. The feature does not introduce end-to-end encryption, change data-retention policies, erase previously saved memories, or modify the setting that allows conversations to be used to improve models.

 

OpenAI also warns that resetting the PIN permanently deletes the locked chats and their associated files and content. Certain conversations, including Temporary Chats, group chats, archived chats, and chats with scheduled tasks, cannot be locked.

 

The feature is rolling out across ChatGPT Free, Go, Plus, Pro, Business, Enterprise, and Edu plans, although availability may vary during the rollout. ChatGPT for Healthcare is excluded.

 

The bigger picture: As people increasingly use AI assistants for personal and sensitive discussions, privacy controls are becoming more important. Chat locking offers greater control over who can casually access conversations on a shared device, but users should still be mindful of the sensitive information they share with AI systems.