6 min read
Most of us have handed a phone to a family member or shared a laptop screen in a meeting and felt a small jolt of worry about what might pop up. If you use ChatGPT for personal questions about health, money, a relationship or a surprise gift, your chat list can say a lot about you at a glance.
This week OpenAI started giving ChatGPT users a way to put selected conversations behind a PIN. The feature is called Locked Chats. It is a useful tool, and it also comes with a few conditions that are easy to skip past. One of them can wipe out everything you locked.
What happened
On Tuesday, October 20, 2026, OpenAI announced Locked Chats and began a gradual rollout, according to TechCrunch and BetaNews. Because the rollout is staged, some people will not see the option yet. TechCrunch, the OpenAI Help Center and Deccan Chronicle all list the same plans: Free, Go, Plus, Pro, Business, Enterprise and Edu. The Help Center says the feature is not offered in ChatGPT for Healthcare.
Here is how it works, based on OpenAI's help page. You open a chat's three-dot menu, choose Lock chat, confirm, and then create or enter a 4-digit account PIN. On the web and in the desktop app the PIN is the way in. On supported phones you can use Face ID, Touch ID or your normal device unlock instead. Once locked, the chat leaves your regular history, your pinned chats and regular search. Locked conversations are gathered under Projects, in a section called Locked chats. BetaNews notes that you can lock a chat before you start it or lock one you already have.
OpenAI describes the purpose plainly. In wording reported by Deccan Chronicle, the company said the feature can help "keep sensitive conversations out of view" when you share your screen or a device. TechCrunch adds that OpenAI mentioned personal talks and surprise planning as examples.
Not every chat can be locked. TechCrunch lists temporary chats, archived chats, group chats, chats opened from a shared link, chats with scheduled tasks and chats with meeting recordings as off limits. OpenAI's help page adds a few more cases, such as chats tied to background tasks.
On October 24, TechRadar published a closer look at the conditions in OpenAI's help page. When I checked on October 25, I found no reports that OpenAI had changed these rules since launch. The help page used here was last updated October 22.
KRUDO's Fine Print
OpenAI is open about the limits. They sit in its help page rather than in the lock screen, so here they are in one place.
- Resetting a forgotten PIN costs you the chats. You can change your PIN and keep your locked chats. Forgetting the PIN deletes nothing by itself, but if you reset it, the Help Center says every locked chat is permanently deleted, along with memory kept only in those chats and files protected only by the old PIN. OpenAI Support cannot bring them back. Workspace administrators cannot reset your PIN for you either.
- Old shared links keep working. If you shared a chat with a link before locking it, the link still shows what was shared. Messages you add after locking are not added to it. You have to delete the link yourself to cut off access.
- Memory you already built stays put. Memory is the ChatGPT feature that saves details about you for later answers. Anything it saved before you locked a chat stays in your regular Memory and can still shape replies. New details from inside locked chats are kept apart and used only in other locked chats, according to TechCrunch and the Help Center.
- A locked chat can stay on screen. On the web, a chat can remain open in the tab where you locked it. Other devices that already had ChatGPT open may keep showing the title until their list refreshes, and on iPhone, older content may linger until the screen updates.
- No added encryption. The Help Center, TechCrunch and BetaNews all say locking adds no end-to-end encryption (scrambling that only you could undo) and no separate encrypted vault.
- Nothing changes on OpenAI's side. On retention and training, OpenAI's help page says locking does not change how long OpenAI keeps the chat and does not change your "Improve the model for everyone" setting. So a locked chat follows the same rules as your other chats. The page does not give a separate retention period for locked chats.
- Work accounts still see logs. In Enterprise and Edu workspaces, the Help Center says locked chats are still included in message logs available through OpenAI's Compliance Platform. Locking also does not override workspace policies or legal duties.
- Anyone who opens it can copy it. The Help Center notes that a person who can open a locked chat can copy what is in it.
How this kind of thing happens
A PIN lock like this is a screen-level gate. Your conversation is still stored on the company's servers like any other chat. The app simply hides it from your lists and asks for the PIN before showing it. That is why the lock protects you from people near your screen but does not change what the company itself holds or how its data rules apply.
The reset rule follows the same pattern. OpenAI's help page says a reset permanently deletes the locked content and that its support team cannot restore it, so remembering the PIN is the only safeguard.
What it means for you
For the everyday problem of a shared tablet, a kid borrowing your phone or a screen share at work, Locked Chats does what it says. It keeps titles and content out of casual view, and it is offered on every ChatGPT plan from Free up, according to OpenAI.
It works like a privacy curtain. If your worry is what is stored about you, the rules are the same as before. If you use ChatGPT through an employer or school account, your organization's logging rules still apply. And the biggest personal risk is self-inflicted: forget the PIN and reset it, and the locked conversations are gone for good.
For a truly one-off sensitive question, OpenAI offers a different tool. Its help page says a Temporary Chat stays out of your history, does not create or update memories, and is not used to improve its models while it stays temporary. OpenAI may keep a copy for up to 30 days for safety reasons. The trade-off, as TechCrunch points out, is that a locked chat stays available to return to. OpenAI's help page says a temporary chat you close without saving cannot be reopened from your history.
What you can do
- Pick a PIN you will remember. Write it down somewhere safe offline. A reset deletes every locked chat, per OpenAI's help page.
- Lock before you start. TechRadar suggests locking a chat before the sensitive part begins, so new details stay out of your regular Memory.
- Clean up old Memory. Details saved before locking stay in regular Memory. Review and delete them by hand in ChatGPT's memory section, as TechRadar advises.
- Delete old shared links. Locking does not revoke them. Remove any link you made for a chat you have since locked.
- Close or refresh tabs, and check other devices. OpenAI recommends checking other open devices before you share your screen or hand over a device.
- Use Temporary Chat for one-off questions. Choose Temporary when you open a new chat. OpenAI says it is not used for model training while temporary.
- Check your training setting. In Settings, then Data controls, you can turn off "Improve the model for everyone", according to OpenAI's Data Controls FAQ.
From the publisher · KRUDO Designs
This part is about our own product.
The free HOLLIS app for Android from KRUDO Designs is built and now in testing before launch. It talks to one place only: the HOLLIS H-Ai in your home. No company chat database. Your conversations stay in your household.
Sources
- TechRadar, Oct 24, 2026
- BetaNews, Oct 21, 2026
- TechCrunch, Oct 20, 2026
- Deccan Chronicle, Oct 21, 2026
- OpenAI Help Center (Using locked chats), Oct 22, 2026
- OpenAI Help Center (Temporary Chat), Sep 26, 2026
- OpenAI Help Center (Data Controls FAQ), Oct 22, 2026
This article is based on the published reports listed above, written in our own words. Details are as reported on the dates shown. Spotted an error? Email support@krudodesigns.net and we will correct it.
