OpenAI has launched a new 'Locked Chats' feature allowing ChatGPT users to protect specific conversations using a PIN, Face ID, or Touch ID, preventing unintended access when sharing devices with others.
- Chats can be locked and accessed via PIN, Face ID, or Touch ID
- Locked Chats have separate memory and don’t affect regular chat history
- Feature rolls out for most ChatGPT plans except Healthcare
What happened
OpenAI announced a new feature called Locked Chats, enabling users to safeguard particular ChatGPT conversations with a four-digit PIN or biometric options such as Face ID and Touch ID on supported devices. This addition aims to give users peace of mind when sharing their screen or device with others.
Once locked, these chats move to a special section distinct from normal conversations, pinned chats, and search results. Unlike Temporary Chats, Locked Chats remain accessible indefinitely and have a separate memory space that does not influence regular conversation contexts.
Why it matters
As ChatGPT becomes more integrated into daily life, handling personal and confidential topics, the need to protect sensitive information on shared devices grows. Locked Chats help address privacy concerns without requiring users to exit the app or clear chat histories.
However, this feature does not implement end-to-end encryption, alter how data is stored, or affect OpenAI’s data retention and model improvement settings. It primarily enhances local access control, leaving broader data privacy considerations unchanged.
What to watch next
The feature is beginning to roll out across most ChatGPT offerings, including Free, Plus, Pro, Business, Enterprise, Go, and Edu plans, but it is not available for Healthcare users or certain chat types like group or archived chats. User adoption and feedback will shape how OpenAI might expand or refine privacy controls.
Future developments could include extending locked chat protection to additional chat categories, enhancing encryption standards, or broader privacy tools as demand grows amid increasing scrutiny of AI data handling practices.