What is ChatGPT's Locked Chats feature and why does it matter?
ChatGPT now lets users protect specific conversations behind a PIN or biometric lock, adding a layer of privacy by segregating those chats from regular history and search. This helps prevent casual access by others who might use your device, making it easier to keep sensitive discussions away from prying eyes. Such protection can be useful if you share your device or don’t want certain chats to appear in your main conversation feed.
How does Locked Chats actually work?
When you lock a chat, ChatGPT moves it into a separate 'project' that requires authentication to access. You set up a four-digit PIN or use your device's face or fingerprint recognition. Locked chats won’t show up in your search or history unless you unlock them. Additionally, new messages in Locked Chats remain isolated from your regular ChatGPT memory, so their contents won’t influence responses in unlocked chats.
Important operational details
- You must remember your PIN; forgetting it and resetting will permanently delete all locked chats tied to the old PIN.
- Simply locking a chat doesn’t close it immediately on open browser tabs or previously logged-in devices; tabs must be refreshed or closed manually.
- After unlocking, the chat remains accessible until you manually lock it again; navigating away doesn’t automatically re-lock the chat.
What privacy limitations should users be aware of?
Despite the Locked Chats feature improving visual and direct access control, it does not fully prevent sensitive information exposure:
- Pre-lock memory leakage: If ChatGPT has already incorporated details from your conversation into its general memory before locking, that information remains accessible within other chats. You must manually delete such memory entries to fully secure prior details.
- Shared link loophole: Any shared URLs to conversations remain accessible even after locking chats. Locking does not revoke access from previously distributed links; these must be deleted separately.
- Temporary visibility: Locked chats can remain visible on other devices until refreshed, so immediate privacy isn’t guaranteed across all logged-in sessions without user action.
When to consider Temporary Chats instead
If your conversation demands stronger confidentiality—no saving, no memory, no sharing—Temporary Chats may be a better choice. These chats don’t appear in history, don’t save memory details, and aren’t stored by OpenAI for model training, but you lose the ability to revisit them later.
Bottom line: What Locked Chats means for ChatGPT privacy
Locked Chats offer a meaningful but partial solution to everyday privacy concerns for ChatGPT users. They effectively guard against casual browsing of your account and device by others, but don’t automatically secure sensitive details already stored in memory or shared externally. To fully protect your information, you need to complement locking with proper memory management and cautious link sharing. For highly sensitive topics, Temporary Chats remain the strongest option. Understanding these trade-offs empowers users to safeguard their ChatGPT interactions more effectively.