OpenAI has added a way to lock individual ChatGPT conversations behind a passcode or biometric check, similar to how you might hide a photo album on your phone. The feature is aimed at everyday privacy moments — keeping a surprise party chat hidden when you hand your phone to a friend, for instance — not at protecting business data.
Here's how it works: users can select a specific chat and lock it, requiring a device passcode, fingerprint, or face scan to reopen it. Locked chats are also hidden from search results and chat history lists within the app. It's an opt-in feature applied chat-by-chat, not an account-wide setting.
This arrives as ChatGPT has become a default tool in many small offices — drafting emails, summarizing contracts, brainstorming marketing copy, even fielding HR questions. As that usage has grown, so has scrutiny of what happens to the information typed into these chats, including concerns about chat histories being used for model training or surfaced in legal discovery requests.
OpenAI frames this specifically as a personal-use convenience, comparable to hiding a text thread, not as a response to business security concerns. The company has separately rolled out data controls for business and enterprise tiers over the past two years, including options to exclude conversations from training data and admin-level retention settings. This new feature sits apart from those tools entirely.
The move also follows a broader pattern among consumer tech companies adding lightweight privacy controls to chat and messaging products — think locked notes apps, hidden folders, or disappearing messages. These features tend to address social privacy, like someone glancing at your screen, rather than the kind of data protection a business needs for client records or financial details.
For small business owners, the distinction matters more than it might seem. A locked chat still lives on OpenAI's servers and is still subject to whatever data retention and usage policies apply to your account tier. Locking a conversation prevents a coworker from scrolling past it on your laptop; it does not encrypt the data, exempt it from potential legal requests, or change what OpenAI can see or use depending on your plan.
If your business handles sensitive client information, medical details, financial records, or anything under confidentiality obligations, this feature doesn't change your compliance posture. Those protections still depend on which ChatGPT plan you're using — consumer, Plus, Team, or Enterprise — since training-data opt-outs and retention controls vary significantly between them. This is a good moment to actually check which tier your business is on, because many small teams default to personal or Plus accounts without realizing the data-handling terms differ from the business tiers.
The practical move this week is a short audit: list what kinds of information employees are pasting into ChatGPT, confirm which account type they're using, and verify whether your plan excludes conversations from model training. If the answer is unclear, that's worth resolving before assuming a locked chat icon means your data is protected.
Watch for whether OpenAI extends locking to shared Team or Enterprise workspaces, and whether competitors like Google or Anthropic introduce comparable consumer-facing privacy toggles. Also worth tracking: whether OpenAI eventually ties locked chats to stricter data handling, which would signal a shift from a cosmetic feature toward an actual security control.
The bottom line: ChatGPT's new lock feature adds a layer of screen privacy, not data security, and small businesses should keep making that distinction when deciding what information belongs in these chats at all.