🔥 ChatGPT Locked Chats: Privacy Feature Explained (2026)
Learn how ChatGPT locked chats work—PIN/biometric gates for ChatGPT privacy feature. Protect conversations, manage AI privacy settings, and understand ChatGPT memory.
ChatGPT Locked Chats: A Privacy Layer, Not a Private Vault
ChatGPT is increasingly used for conversations people wouldn't necessarily want showing up in someone else's view - personal questions, sensitive work notes, or plans for a surprise. Until now, trying to keep one conversation out of sight while you hand your phone to someone else has been clunky.
OpenAI's new Locked Chats feature is meant to solve that specific problem. It lets you protect selected conversations with a four-digit PIN, Face ID, or Touch ID on supported devices.
That's useful, but it helps to be clear about what it does - and what it doesn't. Locked Chats restrict access to a conversation inside the app. It doesn't add end-to-end encryption, change OpenAI's data retention, or alter model-training settings. Think of it as screen-access control, not a new privacy architecture.
This distinction matters if you're deciding whether to lock a personal chat or something you consider confidential at work. Here's how Locked Chats work, what happens to Memory, which conversations are eligible, and where the protection ends.
How ChatGPT Locked Chats work
OpenAI announced Locked Chats on October 20, 2026. The feature lets you put an eligible conversation behind an access check. You can use a four-digit PIN, or Face ID / Touch ID on devices that support it. Once locked, the conversation appears in a dedicated Locked Chats section, separate from the regular chat list, pinned conversations, and chat search results.
Day-to-day, it's straightforward: pick a conversation to lock, choose the available authentication method, then go back to the Locked Chats section when you want to open it again. The conversation is still there for you - it's just not sitting in your normal list. That's different from Temporary Chats, which are a separate session and don't appear in chat history.
The practical difference is the point. A Temporary Chat is for something you don't want included in your everyday history. A Locked Chat is still a conversation you plan to revisit, but you want an extra step so someone else can't open it while using your device.
The use case is device sharing
OpenAI frames Locked Chats around situations like discussing something personal or planning a surprise, then handing your device to a friend or family member so they can see another conversation. In that moment, keeping the sensitive chat out of normal history and search results can lower the odds of an accidental reveal.
It works best for "casual access" risks: someone scrolling the app while borrowing your phone, or a colleague catching your chat list during a screen share. It's not a replacement for securing your phone or account. If someone can unlock your device, access your account, or use an already authenticated session, a chat-level lock should be treated as an extra barrier - not a full security boundary.
Locked Chats are not Temporary Chats
These two options solve different problems:
- Use a Locked Chat when you want to keep a conversation and return to it, while keeping it out of your everyday chat history.
- Use a Temporary Chat when you want a separate session that doesn't show up in chat history.
Neither should be treated as a blanket promise about how information is stored or processed. OpenAI says locking a conversation doesn't change retention or the "Improve the model for everyone" setting. If those details matter to you, check your account's data controls instead of assuming the lock changes them.
What the feature protects - and what it does not
The easiest way to understand Locked Chats is to separate "what's visible in the app" from "what happens to data behind the scenes." Locked Chats are meant to limit whether someone casually using your ChatGPT app can open a selected conversation. OpenAI says they don't turn that conversation into a separately encrypted vault.
According to OpenAI, locking a chat does not:
- Create end-to-end encryption.
- Create a separate encrypted vault for the conversation.
- Change how long the conversation is retained.
- Remove information ChatGPT remembered before the conversation was locked.
- Change the "Improve the model for everyone" setting.
Those limits matter. The lock helps if your worry is someone opening the wrong chat on your device. It doesn't answer the broader question of how the service stores, retains, or might use that information.
A lock is not a data-governance control
In business settings, it's common to mix up interface privacy with information governance. They overlap, but they aren't the same thing. A lock can reduce accidental exposure through the user interface, but it doesn't, by itself, set company policy, control retention, or change account-level data settings.
The same idea applies to individuals. Even a "private" conversation can include information you decided to send to the service. The lock changes how the conversation is accessed in the app; it doesn't rewrite the terms or settings that govern how the service handles data. If the information has contractual, regulatory, or business risk, you'll want the appropriate account controls, along with your organization's approved tools and policies.
For more context, see TechCrunch's report on the announcement. The key point isn't that Locked Chats are useless - it's that they address a narrower risk than the phrase "private chat" might suggest.
PIN reset has a significant consequence
OpenAI says resetting a PIN permanently deletes all Locked Chats, files, and other content protected by the previous PIN. That's more than a small convenience choice. Treat it like a credential you need to manage, and don't reset it if you expect those protected chats or files to stay available.
It also changes how you should think about long-term records. If something is genuinely important to preserve, don't assume the locked area functions like an independent archive or backup. Know what happens when you reset before you put critical content there.
How Locked Chats interact with Memory
Memory is where people can get surprised, so it's worth spelling out. When Memory is enabled, Locked Chats use their own separate memory shared across conversations inside that Locked Chats section. New messages sent after a chat is locked can update that Locked Chats memory, but they don't update regular Memory. OpenAI says this separation is intended to keep new information from locked chats out of regular chats.
In practice, there are two boundaries to keep in mind: information ChatGPT remembered before you locked the conversation, and information added after the chat is locked. Locking doesn't remove earlier remembered information. New information from a locked conversation can contribute to the separate memory associated with Locked Chats, instead of the regular memory used across ordinary chats.
What this means when you switch contexts
Say you lock a conversation that includes personal details, then start a normal chat. The feature is intended to keep new information from the locked conversation out of regular Memory. That's helpful if you want personal follow-ups to stay separate from your day-to-day interactions.
But locking doesn't wipe history after the fact. If ChatGPT had already remembered something before you locked the conversation, locking it doesn't undo that. If your goal is to manage previously saved information, the lock isn't the control you want. Review Memory separately in your account settings.
More broadly, this is a common pattern in AI product design: privacy controls can live at different layers. A conversation might be hidden from lists, excluded from a memory flow, governed by account-level settings, and still be retained by the service under separate rules. One control rarely covers every layer.
A useful question isn't just "Is this chat private?" It's: "Private from whom, for what part of the system, and under which settings?"
For users, that means matching the control to the concern. Use a chat lock to reduce casual access on a shared screen. Check Memory to see what the assistant may recall. Review your account's data settings if your concern is data use or retention. These controls solve different problems, and treating them as interchangeable can lead to avoidable surprises.
Availability and practical use
Locked Chats roll out to users on ChatGPT Free, Go, Plus, Pro, Business, Enterprise, and Edu plans. OpenAI says the feature isn't available on ChatGPT for Healthcare. The rollout started October 20, 2026, so it may not show up immediately for every eligible account.
OpenAI also lists several conversation types that can't be locked:
- Temporary Chats.
- Archived chats.
- Group chats.
- Chats created from a shared link.
- Chats with scheduled tasks.
- Chats with meeting recordings.
If you don't see the option, start by checking whether the conversation falls into one of those excluded categories and whether the feature has reached your account yet. During a rollout, it's normal for interfaces to differ temporarily.
A practical decision checklist
Before you use Locked Chats, ask what you're actually trying to protect:
- Are you mainly worried about someone casually viewing your device? Locked Chats are built for that kind of access-control problem.
- Do you need to return to the conversation? Locking keeps it available; a Temporary Chat serves a different purpose.
- Are you trying to change data retention or model-improvement settings? Locking doesn't change those settings. Use the account controls for that.
- Does the conversation contain information already saved to Memory? Locking won't remove information remembered before the lock. Check Memory separately if needed.
- Can you keep track of the PIN? Resetting it permanently deletes the locked chats, files, and other content protected by the old PIN, according to OpenAI.
For teams, there's one extra question: is ChatGPT an approved place for this information? A UI lock can reduce the chance that a colleague sees a conversation during a device-sharing moment. It doesn't replace organizational rules for customer data, source code, legal material, or regulated information.
Overall, Locked Chats are a reasonable addition because they target a real behavior - people showing one part of an app while other conversations sit in the same screen. The trade-off is that an access control can easily be mistaken for stronger privacy than it actually provides. Use it for what it does, and handle the rest of your privacy decisions separately.
Conclusion
ChatGPT Locked Chats is a focused privacy feature, not a complete privacy upgrade. It helps when you want a conversation out of your normal chat lists and behind an extra barrier before someone else can open it on your device. It doesn't change encryption, retention, existing Memory, or model-improvement settings.
The decision is fairly simple: use the lock to manage casual access inside the app, and use account-level or organizational controls for everything else. Before relying on it, check whether the feature has reached your account, confirm the conversation is eligible, and understand what happens permanently when you reset the PIN.