ChatGPT now lets you lock your chats

OpenAI said on Tuesday that ChatGPT users can now lock individual conversations so that sensitive chats stay out of view when someone else picks up their screen or device. A locked chat can be protected with a four-digit PIN, or with Face ID or Touch ID where the device supports it. The feature is beginning to roll out on Tuesday and may not appear for every user immediately.
Locking is a per-conversation action. Once a chat is locked, it moves into a dedicated Locked Chats section and is kept apart from ordinary chat history, from pinned conversations and from chat search results. That separation is the core of the design: the content does not sit in the main thread list where a friend, family member or colleague scrolling through the app would encounter it.
OpenAI frames the use cases as personal: working through something private, or planning a surprise, while still handing the phone to someone to show them a different conversation. The distinction from the app's existing Temporary Chats matters here. Temporary Chats are a separate session that does not show up in history at all, so there is nothing to come back to. Locked Chats work the other way around — the conversation persists and the user can reopen it whenever they authenticate.
Memory behaves differently inside the section. When Memory is enabled, Locked Chats keeps its own memory that is shared among conversations within that section. Messages sent after a chat is locked can update that locked memory, but they do not write to regular Memory. In practice, that means a detail disclosed in a locked chat should not surface later in an ordinary conversation. The company also says that information ChatGPT had already remembered before a chat was locked stays where it is; locking does not scrub prior memories.
The privacy boundary is narrower than the word "locked" might suggest. OpenAI states that locking a chat does not add end-to-end encryption, does not create a separate encrypted vault, does not change how long the conversation is retained, and does not alter the "Improve the model for everyone" setting. So this is a screen-level access control aimed at other people who handle the device, not a change to how OpenAI itself stores or processes the content. For developers and product teams, that is a familiar split: hiding a record from a UI and encrypting it against the service operator are separate problems, and only the first is being solved here.
The PIN has a destructive failure mode worth flagging to users. OpenAI says resetting a PIN permanently deletes all locked chats, files and other content protected by the previous PIN. Anyone who forgets the code and resets it loses the whole locked collection rather than regaining access to it, which is a deliberate trade-off in favor of not leaving recoverable data behind, but a risky one for people who treat the section as a durable archive.
Eligibility has notable exclusions. Locked Chats cannot be applied to temporary chats, archived chats, group chats, chats created from a shared link, chats with scheduled tasks, or chats with meeting recordings. The feature will be available on ChatGPT Free, Go, Plus, Pro, Business, Enterprise and Edu plans, but not on ChatGPT for Healthcare.
My reading is that the plan mix is the interesting commercial signal: this is being treated as a baseline privacy control rather than a paid upgrade, since it reaches Free and Go as well as the top tiers. The Healthcare carve-out suggests OpenAI is keeping regulated deployments on a separate track, where features that touch how conversation data is surfaced and retained may need different review.
A likely trade-off is friction. Every entry into the section requires a PIN or a biometric check, and the separate memory means users may notice that ChatGPT seems to "forget" things across the boundary between locked and regular chats. That is the intended behavior, but it can read as inconsistency unless the app explains it clearly.
There is also the question of what happens on shared or multi-device accounts, where a locked chat synced to another signed-in device may be reachable with the same PIN. OpenAI's stated protections are about access to the screen and app section, not about account-level compromise, so teams handling confidential work in ChatGPT Business or Enterprise should not treat locking as a substitute for their own data-handling rules.
Why it matters: anyone who uses ChatGPT on a phone that other people occasionally hold gets a way to keep selected conversations out of the main list and out of search, without giving up the chat entirely the way Temporary Chats require. Developers and product teams building comparable assistants should note the pattern — a separate section, its own memory namespace, biometric or PIN gating and destructive PIN reset — as a reusable design for consumer privacy controls. My inference is that this raises expectations for similar features in rival chat products, while the absence of encryption and retention changes means it should not be read as a compliance or confidentiality guarantee.
Based on reporting from the original publisher. Visit the source for full context and later updates.
Publisher excerpt
The company says the feature is designed for situations where users might be discussing something personal or planning a surprise and want to keep those conversations private when handing their device to friends or family to show them something in another chat.