ChatGPT has become a confessional for a lot of people. Health worries, money problems, relationship questions and half-finished work ideas all end up in the same scrolling sidebar, visible to anyone who picks up an unlocked laptop or phone. OpenAI is now giving users a way to tuck those conversations out of sight.
The company has started rolling out locked chats, a protected space that removes selected conversations from the main chat history, pinned chats and search results. Getting back into them requires a four-digit PIN, while mobile users can rely on Face ID, Touch ID or their phone’s standard unlock method instead. The idea will feel familiar to anyone who has used the Locked Folder in Google Photos, WhatsApp’s chat lock or the app-locking tools now common on iPhone and Android.
Availability is unusually broad for a new ChatGPT feature. Locked chats are arriving for Free, Go, Plus, Pro, Business, Enterprise and Edu users across web, mobile and the desktop apps for macOS and Windows. Signs of the feature first surfaced back in August, so this is a fairly quick turnaround from discovery to release.
There are important caveats, and the biggest is about what “private” actually means here. Locked chats are not end-to-end encrypted. This is a privacy screen for your device, designed to stop a partner, colleague or child from stumbling across a conversation, not a guarantee that your data is shielded on OpenAI’s side. Anyone sharing genuinely sensitive information should treat it that way.
The PIN also carries real stakes. OpenAI warns that resetting it wipes every locked chat, and those deleted conversations can’t be recovered. Forgetting a four-digit code could mean losing exactly the conversations you cared most about protecting, so it’s worth choosing something memorable before moving anything important into the vault.
Coverage across ChatGPT’s growing list of modes is uneven, too. Cloud conversations in Chat and Work modes can be locked, but Codex sessions, local Work conversations and ChatGPT for Healthcare chats are excluded. Voice works inside locked chats on mobile, though not on desktop.
The feature reflects how chatbots are drifting from novelty tools into deeply personal ones. As people hand AI assistants more of their private lives, basic device-level protections like this start to look overdue rather than impressive. The more interesting question is whether users will eventually expect the stronger protection that locked chats deliberately don’t offer.
