ChatGPT Locked Chats Roll Out With PIN Protection
OpenAI is rolling out ChatGPT locked chats, with PIN-protected saved conversations and important limits around retention, shared content and PIN resets.
ChatGPT locked chats are rolling out from October 20, 2026, adding PIN-protected access to saved conversations hidden from ordinary history, pinned chats and search. OpenAI’s announcement keeps retention and model-improvement settings unchanged. This is a way to keep a conversation available without leaving it in the usual conversation list. ChatGPT release notes.
ChatGPT locked chats: the rollout and first action
The consumer announcement covers Free, Go, Plus and Pro across web, iOS, Android and the new macOS and Windows desktop app. Open the conversation’s more-options menu and choose Lock chat. Availability is gradual rather than an immediate entitlement on every account.
The detailed help page also lists Business, Enterprise and Edu. Eligible cloud Chat and Work conversations qualify; Codex conversations, local Work conversations and ChatGPT for Healthcare do not. After confirming the lock and supplying a PIN, find the conversation under Projects → Locked chats. Choose Lock now when finished.
Start with a harmless saved conversation rather than the only copy of an important document. Our suggested first check is simply whether the menu appears on the intended account, workspace and client. Record those details if it does not. Changing subscriptions is not a sensible first troubleshooting step for a gradual rollout.
Treat the PIN reset warning as a data-loss decision
OpenAI distinguishes changing a known PIN from resetting a forgotten one. Changing it preserves locked chats; resetting permanently deletes all locked chats, memories held only there and content protected only by the previous PIN. Support cannot recover them.
The locked-chats documentation also explains that locking is not end-to-end encryption or a separate vault. Existing shared-link content remains accessible, and separate Space files retain their access rules. Information remembered before locking is not erased.
For a practical review, write down the outcome you need before choosing a control: conceal a saved conversation, stop sharing an existing copy, change future training use, or remove retained content. Those are four different jobs. A conversation disappearing from the sidebar proves only the visible result you inspected; it is not a receipt for the other three.
Training and temporary chat remain separate choices
OpenAI’s Data controls guide places Improve the model for everyone in Settings → Data controls. Turning it off prevents new conversations from being used for training, but does not hide or delete saved chats. For signed-in users, that preference applies across devices.
Temporary chat has a different purpose. While temporary, a conversation stays outside history, creates no new memories and is not used for model improvement. Existing personalization can be selected before starting; saving the conversation turns it into a regular chat governed by normal account settings.
ChatGPT locked chats therefore fit the “save it, but restrict casual viewing” decision. Temporary chat fits the different decision to avoid a normal saved-history entry. Review the training preference independently of either choice. Our earlier Privacy Center report covers the broader settings navigation rather than this new conversation lock.
Hidden, archived and deleted are different retention states
OpenAI’s chat and file retention guidance says regular and archived chats remain saved until deletion or an applicable workspace retention policy removes them. Archiving does not shorten retention. Deletion removes a chat from account view immediately and schedules system deletion within 30 days, with stated exceptions for de-identified content and security or legal obligations.
Files require their own inspection. A file saved separately in Library can remain after the conversation is deleted. OpenAI instructs users to delete that Library copy separately when they want it removed too.
This matters when evaluating ChatGPT locked chats: a cleaner conversation list is a visibility outcome, whereas retention describes what remains stored. For an illustrative housekeeping checklist, list the conversation, any separately saved document and any shared copy as individual items. Mark each item complete only after inspecting the matching control. Do not infer file deletion from a missing chat title.
Enterprise and Edu logs still include locked conversations
The Enterprise and Edu release notes announce the same gradual rollout across supported web, mobile and new desktop clients. Organization retention policies and compliance requirements continue to apply. Conversation message logs remain available through the OpenAI Compliance Platform.
For a workplace rollout note, the useful distinction is between an employee’s ordinary conversation list and the organization’s governed records. Describe the former as restricted viewing, not as an exemption from the latter. Our recommendation is to pair any employee instructions with the organization’s existing retention policy and approved-content rules, rather than inventing a second policy around a lock icon.
A focused check before using sensitive material
Use the first conversation to confirm the intended client and account, then read the PIN warning before moving important work. Review training use, retained files and existing shares according to the outcome you actually need. If the question is unfamiliar account access rather than conversation visibility, our security-history report explains the separate sign-in review.
ChatGPT locked chats add a useful saved-conversation control. Their value depends on matching that control to its job: keeping selected chats out of ordinary view while preserving the account and workspace rules that still govern the content.