ChatGPT Introduces Locked Chats to Protect Sensitive Conversations

OpenAI has introduced Locked Chats, a new ChatGPT feature that allows users to move selected conversations into a protected section instead of leaving them visible in their regular chat history. The feature, announced on October 20, 2026, lets users secure access with a four-digit PIN or supported device authentication, including Face ID and Touch ID on compatible mobile devices. It is designed for people who want to keep personal conversations out of view when sharing a device or showing someone another ChatGPT conversation.

The new feature adds a layer of access protection without requiring users to delete conversations they want to keep. Locked chats remain saved and can be reopened when needed, but their titles and contents are removed from regular chat history, pinned conversations and standard conversation search. However, OpenAI makes an important distinction: locking a conversation does not create end-to-end encryption, change the company’s data-retention rules or automatically change the user’s model-improvement settings. The feature is rolling out gradually across eligible ChatGPT plans, so it may not appear in every account immediately.

Key Takeaways

• New feature: Locked Chats lets users hide selected conversations from regular ChatGPT history and search.
• Access protection: Users can set a four-digit PIN; supported mobile devices can also use Face ID, Touch ID or device authentication.
• Conversations remain saved: Unlike Temporary Chat, locked conversations can be reopened later.
• Separate section: Locked conversations are collected under a dedicated Locked Chats area.
• Memory separation: New information learned within locked chats does not automatically update regular ChatGPT memory.
• Important limitation: The feature is not end-to-end encryption and does not change how long OpenAI retains conversations.
• Availability: The rollout covers eligible Free, Go, Plus, Pro, Business, Enterprise and Edu users, with some exclusions.

What Are Locked Chats in ChatGPT?

Locked Chats is a privacy feature that separates selected conversations from the standard ChatGPT interface. Instead of deleting a discussion or leaving it visible in the sidebar, users can lock it and require authentication before opening it again.

The feature is intended for situations where someone else might have access to a user’s computer or phone. For example, a person might want to discuss a sensitive personal matter with ChatGPT but still use the same account to ask ordinary questions while a friend or family member is nearby.

Before Locked Chats, users could delete conversations or use Temporary Chat for certain discussions. Those options serve different purposes. Deleting a conversation removes it from the account’s accessible history, while Temporary Chat is intended for conversations that should not appear in regular history or create memories in the usual way.

Locked Chats offers another option: preserve a conversation while making it less visible and requiring a separate authentication step to access it.

How the feature works

Once a user locks an eligible conversation, it moves into a dedicated section called Locked Chats. The conversation no longer appears in regular history, pinned chats or standard conversation search.

To access it again, the user opens the Locked Chats area and enters the account PIN or uses a supported authentication method. On the web and desktop, the PIN is the primary access method. Supported mobile devices can offer biometric or device authentication as well.

This arrangement can help prevent casual viewing by someone who has access to an already-open ChatGPT account. It is especially relevant on shared devices, during screen sharing or when someone briefly borrows a phone.

However, the feature should not be understood as a replacement for device security. Anyone who knows the PIN or can authenticate through an authorised method may be able to access the locked conversations.

How to Lock a Chat in ChatGPT

The feature is being rolled out gradually, so the option may not yet be available in every account. If it has reached yours, the process is straightforward.

1. Open ChatGPT on the web, desktop app or supported mobile app.
2. Find the conversation you want to protect.
3. Open the conversation’s three-dot menu.
4. Select Lock chat and confirm the action.
5. Create or enter your four-digit account PIN when prompted.
6. Open the Locked Chats section whenever you need to return to the conversation.

After locking a chat, its title and content should disappear from regular history and standard search. It may take some time for conversation lists and search results to refresh across devices.

When you finish using a locked conversation, use the available lock control to end access and hide its titles again. On the web, access is associated with the browser tab where authentication took place; refreshing or closing that tab ends access. Other devices may need to refresh their conversation lists before the updated state appears.

If the Lock chat option is missing, the conversation type or platform may not be eligible, or the rollout may not have reached the account yet.

PIN Protection and Biometric Authentication

Locked Chats uses a four-digit PIN to restrict access. On compatible iPhones and Android devices, users can also authenticate with Face ID, Touch ID or their device’s supported unlock method.

The goal is convenience as well as privacy. A PIN can be used across supported web and desktop experiences, while biometrics may make reopening a protected conversation easier on mobile devices.

Users should choose a PIN they can remember but that other people are unlikely to guess. Because the PIN protects access to the locked section, it should not be shared casually.

There is also a significant recovery warning. OpenAI says resetting the account PIN permanently deletes all locked chats and the content protected by the previous PIN, including relevant files and information stored only within locked chats. The deletion cannot be undone, and OpenAI Support cannot recover the deleted content.

Users should therefore treat PIN recovery carefully. A locked conversation is not simply an ordinary chat with a reversible visibility toggle; the PIN is tied to access to the protected content.

Locked Chats Is Not the Same as End-to-End Encryption

The most important limitation is that Locked Chats does not create an end-to-end encrypted conversation or a separate encrypted vault.

The feature changes where conversations appear and how users access them. It is primarily intended to prevent someone casually browsing an already-open account from seeing selected discussions.

That is different from end-to-end encryption, in which message content is designed to remain inaccessible to the service provider as it travels between participants. OpenAI has not described Locked Chats as providing that protection.

The feature also does not automatically change how long OpenAI retains the conversation or whether the account’s existing model-improvement setting applies. Users should continue to review their data controls and relevant privacy settings separately.

This distinction matters because the word “locked” can imply stronger security than the feature actually promises. A conversation hidden from the sidebar is not necessarily protected against every possible form of account access, device compromise or authorised data access.

For sensitive information, users should continue to secure their devices, protect their accounts and avoid sharing information they do not need to disclose.

How Locked Chats Handles ChatGPT Memory

ChatGPT Memory allows the assistant to use relevant information from previous conversations to personalise later responses. Locked Chats introduces a separate memory behaviour for conversations kept in the protected section.

When Memory is enabled, a locked conversation can use relevant information already present in regular Memory. New information from messages sent after the conversation is locked can be remembered for use within other locked chats, but does not automatically update regular Memory.

This separation is intended to prevent information introduced in a locked conversation from being carried into ordinary chats through newly created memories.

However, locking a chat does not remove information that ChatGPT already remembered before the conversation was locked. Users who want to change existing memories must manage those separately through the available Memory settings.

The feature therefore creates a distinction between new memories associated with locked conversations and information already stored in the user’s regular memory. It should not be treated as a way to erase all previous personalisation data connected to a topic.

Locked Chats vs Temporary Chats

Locked Chats and Temporary Chat serve different needs. One is designed to preserve a conversation behind an access control, while the other provides a separate conversation mode that does not appear in regular history.

Feature	Locked Chats	Temporary Chat
Appears in regular history	No	No
Can be reopened later	Yes, after authentication	Not as a normal saved conversation
Requires a PIN or device authentication	Yes, to access locked chats	No separate locked-chat PIN
Creates new regular Memory	No, while locked	No
Changes data-retention rules automatically	No	Follows Temporary Chat’s separate handling rules
Primary purpose	Hide saved conversations from casual viewing	Have a temporary conversation outside normal history

Temporary Chat is more suitable when a user wants a conversation to stay out of normal history and avoid creating regular memories. Locked Chats is intended for conversations that should remain saved but protected from casual access.

Neither option should be interpreted as a guarantee that information is never retained under any circumstances. Users should consult OpenAI’s current privacy documentation for the applicable retention rules and exceptions.

Which Chats Cannot Be Locked?

Not every conversation type supports the new feature. OpenAI’s guidance excludes several categories, including Temporary Chats, archived chats, group chats, conversations created from shared links, chats with scheduled tasks and chats with meeting recordings.

Codex conversations and local Work conversations are also excluded from the general Locked Chats functionality described in OpenAI’s guidance. Availability can differ by platform and workspace configuration.

This means users may need to check whether a conversation is eligible before planning to protect it with a PIN. If the Lock chat option does not appear in the menu, the chat may be ineligible or the feature may not yet be available on that device.

The exclusions are important for people using ChatGPT for a mixture of ordinary conversations, collaborative work and scheduled activities. Locking a chat should not be assumed to preserve every function available in a regular conversation.

Who Can Use Locked Chats?

OpenAI says Locked Chats is rolling out to users on the following plans:

• ChatGPT Free
• ChatGPT Go
• ChatGPT Plus
• ChatGPT Pro
• ChatGPT Business
• ChatGPT Enterprise
• ChatGPT Edu

The feature is not available on ChatGPT for Healthcare, according to the published guidance. Some conversation types remain ineligible regardless of plan.

For enterprise and education environments, locked conversations remain subject to applicable workspace policies, retention settings and compliance obligations. The feature does not override organisational rules or legal requirements that apply to account data.

The rollout is gradual, so a user may not see the feature at the same time as someone else using the same plan. Checking the conversation menu and the latest official help documentation is the most reliable way to confirm availability.

What the Feature Means for AI Privacy

As people use AI assistants for increasingly personal tasks, privacy controls are becoming an important part of product design. Users may ask ChatGPT to help organise private plans, think through sensitive conversations, draft personal messages or explore questions they do not want immediately visible to someone looking at their screen.

Locked Chats addresses one narrow but practical concern: the visibility of saved conversations when someone else can access the user’s device or account interface.

It does not solve every privacy problem. Account security, data retention, model-improvement settings, workplace access and legal obligations remain separate considerations. A user who needs strong confidentiality should understand those limits rather than relying only on a hidden chat section.

The feature’s value will ultimately depend on how reliably it protects against casual viewing, how clearly the interface communicates its limitations and whether users understand the difference between hiding a conversation and encrypting it.

The Bigger Picture

Locked Chats gives ChatGPT users more control over the visibility of selected conversations without requiring them to delete discussions they want to preserve. The PIN and supported biometric options can be useful on shared devices, while the separate memory behaviour reduces the chance that new information from a locked chat will automatically appear in regular conversations through Memory.

The feature also has clear boundaries. It does not provide end-to-end encryption, change retention settings or override workspace compliance obligations. Its arrival reflects a broader shift in AI product design: privacy controls need to address not only how information is stored, but also how easily other people can see it in everyday use.

Looking Ahead

OpenAI is gradually rolling out Locked Chats across eligible plans and platforms. Users who do not yet see the option may need to wait, while those who can access it should review the PIN and recovery implications before moving important conversations into the protected section. The feature is best understood as a practical access-control measure rather than a comprehensive security system.

As AI assistants become more integrated into personal and professional life, users are likely to expect increasingly precise control over conversation visibility, memory and data handling. Locked Chats is a step toward that control, but its effectiveness depends on users understanding what it protects—and what remains governed by their existing account settings and OpenAI’s policies.

Frequently Asked Questions

1. What are Locked Chats in ChatGPT?

Locked Chats lets users move selected conversations out of regular chat history, pinned chats and standard search. The conversations remain saved and can be reopened after authentication.

2. How do I lock a ChatGPT conversation?

Open the conversation’s three-dot menu, select Lock chat, confirm and set or enter your four-digit PIN. The option may not yet be available to every account or every conversation type.

3. Can I use Face ID or Touch ID?

Yes, on supported mobile devices. On web and desktop, access uses the account PIN.

4. Are Locked Chats end-to-end encrypted?

No. Locking a chat changes its visibility and access process, but it does not create end-to-end encryption or a separate encrypted vault.

5. Does locking a chat stop ChatGPT from remembering information?

New information from messages sent while a chat is locked does not automatically update regular Memory. However, the locked chat can use relevant existing memories, and locking it does not remove information already remembered.

6. What happens if I reset my Locked Chats PIN?

OpenAI says resetting the PIN permanently deletes all locked chats and content protected by the previous PIN. This deletion cannot be undone, and Support cannot recover the deleted content.