OpenAI has added a Locked Chats privacy feature to ChatGPT, allowing users to secure individual conversations behind a dedicated PIN code. The tool hides sensitive threads from regular chat history and search bars to prevent accidental exposure during screen sharing. Rollout is underway across Free, Go, Plus, Pro, Business, Enterprise, and Edu accounts.
Activating the privacy lock requires clicking the 3 dots menu next to any conversation title and selecting the Lock chat option. The thread then transfers into a separate area under Projects marked as Locked chats. Accessing these conversations requires entering a personal PIN code, though mobile users can substitute biometric methods like Face ID or Touch ID. OpenAI binds this security code to the master account rather than an individual workspace, meaning separate profiles require individual PIN setup.
Securing a chat removes the text and any uploaded documents from global search results across the application. However, locking a conversation does not provide end to end encryption. Data retention timelines and model training permissions remain governed by the existing account settings. A major risk involves credential recovery. Forgetting the PIN triggers a permanent reset that completely erases all locked chats, unique memories stored within them, and attached files, with customer support unable to retrieve the lost data.
The update is currently deploying to desktop software on Windows and macOS alongside the official mobile applications. Users across all subscription tiers, from free accounts to corporate teams, will receive access as the release reaches global accounts.
