OpenAI adds passcode and biometric locks to shield sensitive ChatGPT conversations

OpenAI has introduced a locked chats feature designed to keep sensitive conversations hidden from anyone viewing a user's screen or handling their device. Available across Free, Go, Plus, Pro, Business, Enterprise, and Edu plans, the update allows users to secure selected threads behind a four-digit PIN, with options for Face ID or Touch ID authentication on mobile devices. Locked exchanges are hidden from regular history, pinned lists, and search results, and are stored in a dedicated section. [1][2]

The privacy layer also segregates memory so details learned in locked discussions do not leak into regular chats. However, the system has strict boundaries. Resetting a forgotten PIN permanently deletes all locked chats and associated files without recovery options. The capability does not work with temporary, archived, group, or healthcare chats, nor with meeting recordings, shared links, or Codex conversations. [1][2]

OpenAI emphasizes that locked chats are intended solely for local physical privacy and do not provide end-to-end encryption. The tool will not alter how long OpenAI stores data, affect model training settings, or bypass legal disclosure requirements. [1][2]