- OpenAI switched on locked chats in ChatGPT on Tuesday, hiding conversations behind a four-digit PIN or Face ID and Touch ID.
- Resetting a forgotten PIN permanently deletes every locked chat, and OpenAI Support cannot recover the content.
- Locking adds no end-to-end encryption and does not change how long OpenAI keeps the chat.
OpenAI added locked chats to ChatGPT today. With a four-digit PIN, users can protect private conversations.
If a user resets a forgotten PIN, all of their locked chats, files, and memory are erased and can’t be recovered.
OpenAI Support and workspace admins cannot bring deleted content back
The OpenAI help page says that forgetting the PIN does not delete anything by itself. Resetting it does.
The company says that a reset “permanently deletes all locked chats,” as well as any remembered information and any files that were only protected by the old PIN. This change can’t be undone, and OpenAI Support can’t get the information back.
There is also no administrator back-up. An admin of a workspace can’t change a user’s PIN or open a locked chat for them.
On the OpenAI Compliance Platform for ChatGPT Enterprise and Edu workspaces, locked chats can still be seen in the message logs. Since the PIN is linked to the OpenAI account and not the workspace, a person who has two accounts will have two PINs.
It’s not the same thing to change a PIN. Anyone who knows the old PIN can get a new one and keep all the conversations.
The current PIN is still needed to make the change, even if the device is set up with Face ID or Touch ID. If a user gets it wrong too many times, ChatGPT will block access until the retry period runs out.
Hidden conversations get no end-to-end encryption
The chats can be found in a section called “Locked chats” under “Projects.”
To access the hidden conversations, users must have the PIN or, on phones that support them, Face ID, Touch ID, or the device unlock method. OpenAI suggested that the tool be used to plan a surprise or give a phone to a family member or friend.
Temporary chats are not saved in history, but locked chats are. Once users have been authenticated, they can get back in at any time.
There is a separate memory for locked chats. New messages in a locked chat can change a memory that is only shared between locked chats.
This way, the new information doesn’t get into normal conversations.
OpenAI said that locking a conversation does not provide end-to-end encryption or a separate encrypted vault.
It doesn’t change the “Improve the model for everyone” setting, and OpenAI keeps the chat for the same amount of time.
Also, it doesn’t delete anything that ChatGPT had remembered before the lock.
OpenAI says that locked chats could also show ads to users whose accounts get ads. Cryptopolitan has already reported that the company plans to test display ads this month while ChatGPT images are being made.
ChatGPT Free, Go, Plus, Pro, Business, Enterprise, and Edu all have locked chats. ChatGPT for Healthcare is not included.
It is not possible to lock temporary, archived, or group chats. It’s also not possible to lock chats started from a shared link, scheduled task chats, meeting recording chats, or Codex conversations.
Since the feature is rolling out slowly, the Lock chat option may not yet be available on all accounts.
Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.
FAQs
What happens if I forget my ChatGPT locked-chats PIN?
Forgetting it deletes nothing, but resetting it permanently erases all locked chats, and OpenAI Support cannot recover them.
Which ChatGPT plans support locked chats?
Free, Go, Plus, Pro, Business, Enterprise and Edu support locked chats, but ChatGPT for Healthcare does not.
Does locking a chat encrypt it?
No. Locking adds no end-to-end encryption or separate encrypted vault, OpenAI says.
Disclaimer. The information provided is not trading advice. Cryptopolitan.com holds no liability for any investments made based on the information provided on this page. We strongly recommend independent research and/or consultation with a qualified professional before making any investment decisions.
Randa Moses
Randa Moses is an editor and reporter at Cryptopolitan covering tech, AI, robotics, crypto, scams, and hacks. She has worked in the crypto space since 2017. She held roles at Forward Protocol, AmaZix, and Cryptosomniac. Randa holds a degree in Electrical and Electronics Engineering from the University of Bradford.
