ChatGPT's new Locked Chats feature has a major privacy catch

ChatGPT's latest Locked Chats feature introduces a privacy concern that warrants attention. The new functionality locks specific conversations behind a four-digit PIN or biometric authentication on mobile devices. OpenAI implemented this feature to prevent sensitive information from being inadvertently discovered by friends or family members.

To utilize Locked Chats, users must first set a PIN or associate it with their device's face or fingerprint. It's essential to remember the PIN, as losing it results in the permanent deletion of all locked chats and any protected data.

While Locked Chats offer a degree of privacy, there are notable caveats. If a chat remains open in the browser tab after being locked, previously opened devices might still display the chat's title until they refresh. Furthermore, conversations that have already been shared cannot be locked retrospectively. Locked chats do not erase any information previously stored in ChatGPT's memory, so any details shared before locking will still influence the AI's responses.

Users must manually remove that data from the memory section. In situations where maximum privacy is paramount, Temporary Chat is a more suitable option, as it does not appear in regular history, retain memories, or be saved for OpenAI's model improvement program.