ChatGPT Introduces Locked Chats to Hide Sensitive Conversations Behind PIN-Protected Vault

OpenAI Rolls Out Locked Chats Feature for ChatGPT

OpenAI announced a new privacy feature for ChatGPT designed to protect sensitive chats from prying eyes. Called Locked Chats, this feature addresses a common concern among users who access ChatGPT in crowded spaces like buses or subway platforms, or when sharing devices with friends and family1

1

. The feature is rolling out gradually across ChatGPT Free, Go, Plus, Pro, Business, Enterprise, and Edu plans, with availability extending to macOS and Windows desktop platforms2

2

.

Locked Chats moves selected conversations out of regular chat history, pinned chats, and conversation search results into a protected section accessible only through authentication1

1

. Users can find these conversations under Projects > Locked chats in the ChatGPT sidebar, where titles remain hidden from view2

2

.

How the PIN-Protected Vault Works

To hide sensitive conversations, users must create a 4-digit PIN that works across all devices when signed in. On mobile devices, the feature supports biometric authentication including Face ID, Touch ID, or device screen lock methods3

3

. Locking a chat requires opening the conversation, tapping the three-dot menu, selecting "Lock chat," and confirming with the PIN2

2

.

OpenAI implemented memory segmentation to prevent information leakage between locked and regular chats. With Memory enabled, locked chats maintain their own separate memory shared only between conversations inside the protected vault. New messages sent in locked chats update only that isolated memory, ensuring information from locked conversations doesn't leak into regular chats1

1

. However, locking a chat doesn't remove information ChatGPT remembered before the conversation was locked2

2

.

Critical Limitations and Security Considerations

Despite offering an additional privacy layer, Locked Chats doesn't deploy end-to-end encryption. OpenAI can still access locked conversations and potentially use them for model training if users have opted into the "Improve the model for everyone" setting1

1

. The feature doesn't change OpenAI's existing chat retention policies or override applicable workspace policies, safety requirements, or legal and compliance obligations2

2

.

A significant restriction involves PIN management. Users who forget their PIN and choose to reset it will permanently lose all locked chats and protected content. OpenAI emphasizes this as a critical safeguard, though users who remember their current PIN can change it without losing any data3

3

. Additionally, certain conversation types cannot be locked, including temporary chats, archived chats, group chats, and chats created from shared links2

2

.

What This Means for ChatGPT Users

The introduction of Locked Chats reflects growing demand for privacy controls in AI tools used across personal and professional contexts. For Enterprise and Edu workspaces, locked chats remain subject to organizational policies and compliance requirements, balancing individual privacy with institutional oversight2

2

. As the feature rolls out gradually, users should watch for the "Lock chat" option appearing in their conversation menus. The biometric scan option on supported devices offers convenient access while maintaining security, particularly valuable for mobile users who frequently access ChatGPT on the go1

1

.