Signals/ ChatGPT & OpenAI/ Brief /
ChatGPT can now hide chats behind a PIN, but it isn't encryption
OpenAI's new Locked Chats keep sensitive conversations out of history and search. Your data settings stay the same.
OpenAI began rolling out Locked Chats in ChatGPT on October 20. You pick a conversation, lock it, and it disappears from your regular history, pinned chats and search. To open it again you use a four-digit account PIN, or Face ID, Touch ID or your device unlock on supported phones.
Coverage lists Free, Go, Plus, Pro, Business, Enterprise and Edu plans, though the rollout is gradual. Locked chats get their own Memory, which does not carry over to your regular chats. OpenAI pitches the feature at screen sharing and shared devices.
Some chat types can’t be locked, including temporary chats, archived chats, group chats and chats with shared links.
The catch: This is not end-to-end encryption. Locking changes nothing about how long OpenAI keeps chats or whether they help train models. And if you reset a forgotten PIN, every locked chat and its files are deleted for good.
Try this prompt
Use a locked chat for private planning. Lock it first, then paste:
Act as a private advisor. I want to think through [a sensitive topic, e.g. a salary negotiation]. Ask me five questions one at a time, then summarize my options in a short list.
What to do now
- Write your PIN down somewhere safe before you lock anything.
- Delete any shared link for a chat separately. Links made before you lock it stay active.
- If privacy from OpenAI matters, check your data settings too.