ChatGPT Locked Chats: How to Use Them and What They Don't Protect

OpenAI began rolling out Locked Chats in ChatGPT on October 20, 2026. You can lock an individual conversation behind a 4-digit account PIN, or Face ID, Touch ID or device unlock on supported phones, and it moves into a hidden "Locked chats" area. The feature hides chats from people looking at your screen. It does not encrypt them, change retention, or alter model-training settings.

What are ChatGPT Locked Chats?

Locked Chats is a privacy setting that puts individual ChatGPT conversations behind a PIN, so nobody using or watching your device can casually open them. TechCrunch reported that OpenAI framed it around everyday situations, such as handing a phone to family or keeping a gift idea secret.

According to OpenAI's help center guide to locked chats, a locked conversation:

• Moves into a dedicated Locked chats area, under Projects
• Disappears from regular chat history, pinned chats and conversation search (search results can take a while to update)
• Leaves any project it was in, along with that project's instructions and files

In August, TechRadar reported that Android Authority had found references to the feature in ChatGPT's Android app code.

Who can use Locked Chats, and on which devices?

Locked Chats is available on ChatGPT Free, Go, Plus, Pro, Business, Enterprise and Edu plans, but not on ChatGPT for Healthcare. OpenAI's help page lists web, iOS, Android and the new ChatGPT desktop apps for macOS and Windows as supported platforms.

Because the rollout is gradual, some accounts won't have it yet. If you don't see Lock chat in a conversation's ••• menu, it hasn't reached your account. Codex conversations can't be locked.

How do you lock a chat in ChatGPT?

To lock a chat, open its options menu and choose Lock chat. These steps come from OpenAI's help center:

1. Open the ••• menu on the conversation you want to protect.
2. Select Lock chat, then select Lock chat again to confirm.
3. The first time, ChatGPT asks you to create a 4-digit PIN and enter it twice. After that, you just enter it.
4. To reopen a locked chat later, go to Projects, select Locked chats, and enter your PIN or use Face ID, Touch ID or device unlock.
5. To start a new private conversation, open Locked chats the same way and type a message there.
6. To unlock a chat for good, open it, choose •••, select Unlock chat and confirm.

On web and desktop, access relocks when you leave the tab or window. On mobile, it ends after one minute away.

How does memory work with locked chats?

If Memory is on, locked chats get their own memory that never flows back into regular chats, though they can still draw on your regular Memory.

Two details matter:

• Locking does not erase old memories. Anything ChatGPT remembered before you locked the chat stays in regular Memory.
• Unlocking can expose the conversation. After a permanent unlock, the full chat may feed regular Memory, if your settings and workspace policies allow it.

If your organization has turned Memory off, locked chats can't use or create memories.

What don't Locked Chats do?

Locked Chats limits who can see a conversation on your device. It does not change what OpenAI stores. OpenAI's own documentation says locking does not:

• Add end-to-end encryption or create a separate encrypted vault
• Change how long OpenAI keeps the chat
• Change your "Improve the model for everyone" setting
• Remove information ChatGPT already remembered
• Override workspace policies or legal and compliance obligations

Some conversations can't be locked at all: temporary, archived and group chats, chats started from a shared link, and chats with scheduled tasks, meeting recordings or background tasks.

Smaller catches: ads can still appear inside locked chats, and locking a chat with an existing shared link leaves that link working.

The biggest catch is the PIN reset. OpenAI warns that resetting a forgotten PIN permanently deletes every locked chat, any memory stored only in locked chats, and files protected by the old PIN. Support can't recover them, and workspace admins can't unlock your chats or reset your PIN.

What do privacy advocates say?

Privacy groups have mostly focused on the gap between hiding chats and protecting them. TechCrunch noted that locking restricts access but doesn't change how sensitive data is stored or handled. That's the issue the Electronic Frontier Foundation has been pushing on.

In December 2025, the EFF argued that chatbot companies should protect conversations from bulk surveillance, asking them to require warrants, fight broad data orders and publish transparency reports. Its AI and encryption page adds that AI features shouldn't weaken encrypted communications. A screen-level PIN does nothing about server-side access.

OpenAI's leadership has said publicly that more is needed. In August 2025, Axios reported that Sam Altman said OpenAI was "strongly considering" encryption, possibly starting with temporary chats, with no timeline. The same report noted a May 2025 federal court order requiring OpenAI to retain temporary and deleted chats. Locked Chats does not deliver the encryption Altman described.

The fair reading: it fixes a real problem (shoulder-surfing, shared laptops, screen sharing) but isn't a confidentiality guarantee. In Enterprise and Edu workspaces, locked chats still appear in OpenAI Compliance Platform logs.

What does this mean for marketing teams?

For marketing teams, Locked Chats helps with screen privacy but doesn't fix data hygiene. Teams paste unreleased briefs, client positioning, pricing and media plans into ChatGPT. A lock keeps that off a shared screen during a client call, but doesn't change where the data lives or who can audit it.

A few practical rules hold up:

• Decide what never goes into a consumer AI tool, locked or not. Client contracts, personal data and embargoed launches usually belong in that bucket.
• Use the right plan for client work. Business and Enterprise plans carry workspace controls that personal accounts don't.
• Write down the PIN reset risk. If a strategist resets a forgotten PIN, the locked work is gone and IT can't recover it.
• Treat locking as a display setting, the same way you'd treat a privacy filter on a laptop.

This sits next to a wider question we cover often: what AI systems learn and say about your brand. Our notes on AI brand intelligence and how ads now appear inside ChatGPT are useful background, and Model Share tracks how models describe your brand compared with competitors. Good data hygiene on the input side and visibility on the output side work best together.

FAQ

Are ChatGPT Locked Chats encrypted?

No. OpenAI says locking does not add end-to-end encryption or a separate encrypted vault. It only controls access on your devices.

What happens if I forget my Locked Chats PIN?

Nothing is deleted just because you forget it. Resetting the PIN, however, permanently deletes all locked chats, locked-only memories and protected files, and OpenAI can't recover them.

Does locking a chat stop OpenAI from training on it?

No. Locked chats follow your existing "Improve the model for everyone" setting. To keep a chat out of training, change that setting or use a temporary chat.

Can my company admin see my locked chats?

Admins can't unlock your chats or turn the feature off. However, locked chats don't override workspace data access, and in Enterprise and Edu they appear in Compliance Platform message logs.

Is Locked Chats the same as a temporary chat?

No. Temporary chats stay out of your history and aren't used for training, and you can't reopen them later. Locked chats are saved and can be reopened.

Last updated October 22, 2026.