TL;DR: OpenAI has begun rolling out a feature that lets ChatGPT users lock individual conversations behind a PIN or biometric scan, so a sensitive chat doesn’t just sit there in plain view to anyone who picks up an unlocked phone.
It’s a small feature with an obvious use case, and the surprising part is that it’s taken this long to arrive. ChatGPT conversations can contain medical questions, financial details, draft messages to an ex, legal worries, or anything else a person wouldn’t necessarily want visible if they handed their phone to a friend to show a photo and the chat history happened to be the last thing open. Locking a chat is standard practice in messaging apps already. ChatGPT is just catching up.
How it actually works
The rollout, first spotted through hints in the Android app’s code before surfacing more broadly, lets a user lock a specific conversation individually rather than locking the entire app. That distinction matters more than it might seem. A single app-wide lock protects everything behind one gate, which is fine for privacy but annoying if you’re constantly dipping in and out of ChatGPT for mundane tasks and don’t want a PIN prompt every single time. Per-conversation locking lets someone keep most of their chat history easily accessible while sealing off the specific threads that actually need protecting.
Unlocking uses whatever authentication the device already supports for sensitive apps, a PIN or the phone’s built-in biometric options like a fingerprint or face scan, rather than requiring a separate password specific to OpenAI’s app. That’s the right call from a usability standpoint; asking people to remember yet another password for a chat app would guarantee most people just skip the feature entirely.
- Granularity: individual conversations, not the whole app
- Authentication: device PIN or biometric, no separate password required
- Rollout: appears to be staged, starting on Android before a wider release
- Why now: ChatGPT’s user base has grown well past the early-adopter crowd, and ordinary people have ordinary privacy concerns about what’s sitting in their chat history
Why this took until late 2026 to ship
OpenAI’s community forums have had feature requests for exactly this since at least last year, which is a reminder that even well-resourced AI companies prioritize headline model capabilities over what looks, from the outside, like a straightforward quality-of-life feature. That’s not necessarily poor judgment on OpenAI’s part. A locked-chat feature doesn’t move any benchmark or win any comparison against Gemini or Claude. It also doesn’t generate a press cycle the way a new model release does. It’s the kind of feature that matters enormously to the people who need it and is invisible to everyone else, which makes it easy to deprioritize right up until enough users complain loudly enough.
The timing also lines up with ChatGPT’s continued push into more personal, more agentic use cases, memory, scheduled tasks, and deeper integration with a user’s actual life and accounts. The more ChatGPT knows about someone, and the more it can act on their behalf, the more that chat history starts looking less like search queries and more like a diary, and diaries are exactly the kind of thing people want a lock on.
Anthropic’s own assistant has been moving in a similarly personal direction with its recently opened Claude Duet research preview, which leans into proactive, ongoing assistance rather than one-off questions, a trend that makes privacy controls like chat locking feel less like a nice-to-have and more like table stakes across the whole category.
What this doesn’t solve
A locked chat on your own device doesn’t change what OpenAI itself can see or how long that data is retained on its servers, which is a separate and more complicated question than a PIN screen can address. It also doesn’t protect against someone who has your device unlocked and your biometrics already enrolled, which covers a lot of the realistic household scenarios people actually worry about. It’s a real improvement over nothing, not a comprehensive privacy solution, and users should understand that distinction before assuming a locked chat means the conversation is somehow more private from OpenAI itself.
Everyone else figured this out a while ago
WhatsApp has supported locking individual chats behind biometric authentication for a few years now. Apple’s own Messages app lets users hide specific conversations on iOS. Signal, built from the ground up around privacy as the entire product thesis, has offered screen-lock and per-chat protections since relatively early in its life. Against that backdrop, ChatGPT catching up isn’t really news about ChatGPT specifically so much as it’s a marker of how thoroughly an AI chat app has become, for a lot of people, just another messaging app that happens to talk back, with all the same ordinary privacy expectations that come with that category.
That shift in how people relate to ChatGPT, less like a search engine, more like a messaging app holding genuinely personal conversations, is probably the bigger story underneath this one small feature. Features tend to arrive once a product’s actual usage pattern demands them, not before, and OpenAI shipping this now is as much a signal about how people are actually using ChatGPT today as it is about the feature itself.
The bottom line
This is an overdue, welcome, genuinely useful feature that solves a real everyday problem without pretending to solve a bigger one it can’t. Turn it on for anything you’d rather a houseguest not stumble into, and don’t mistake it for data privacy from OpenAI’s own servers.
The Teck Hustlers Editorial Team covers technology, business, finance and gaming news for readers who want the story behind the headline. Every article is researched, written and edited by our in-house team, checked against company announcements, filings and reputable reporting, and updated when new facts emerge. We focus on what each development means for founders, professionals, investors and everyday users. Have a tip or a correction? Reach us through our contact page.
