OpenAI has introduced a new ChatGPT privacy option: Locked Chats, which lets you hide specific conversations with extra protection.
With this new option, you can protect private conversations using a four-digit PIN. Locked chats won’t show up in your regular history, pinned chats, or search results. They are kept in a separate Locked chats section. If you use iOS or Android, you can use Face ID or other biometric systems instead of entering the PIN every time.
This update makes sense as ChatGPT becomes a more advanced tool. For example, OpenAI recently launched an interactive Intelligent UI for new models. Now, your AI experiments are better protected from anyone who might use your phone.
Even with a “lock,” this doesn’t confirm privacy. OpenAI makes it clear that the feature doesn’t offer end-to-end encryption. Your data is still stored on the company’s servers and used to train the model. This is standard across the industry: as of October 2026, 7 of the 9 top chatbots collect user data by default.
This is especially important now that the European Union started fully enforcing the AI Act in August 2026. Still, chat locking only protects your conversations from people around you, not from the company or legal requirements.
Keep in mind that if you forget your PIN and reset it, OpenAI will delete all locked chats and any files uploaded there. Support won’t be able to recover them, so remember your PIN or use biometrics.
The feature is rolling out gradually and will be available to both free users and Go, Plus, and Pro subscribers. On mobile devices, locked chats support voice mode, which is handy for people who use AI while on the move. This can also help in professional settings, like drug development with AI agents, where keeping information secure is crucial. However, you can’t lock dialogues in Codex yet.