OpenAI has rolled out a new ChatGPT feature that lets you tuck away your more sensitive conversations behind a lock. It’s called Locked Chats, and it’s the first time I’ve seen this kind of protection on a major AI chatbot. The idea is pretty straightforward, as is clearly evident by the name.
When you move an existing conversation into the Locked Chats section, it is protected by a PIN when you use ChatGPT in a web browser or through the desktop app. On phones and tablets, the protection goes a step further. Besides your device passcode, you can unlock these chats with biometrics, such as Face ID on an iPhone or a fingerprint sensor on Android devices.
The more meaningful part is what happens behind the scenes. When you start a new conversation in Locked Chats mode, nothing from it is used to update ChatGPT’s memory. If you’re not familiar with it, memory lets ChatGPT retain context from your earlier conversations. When you bring up a topic you’ve discussed before, the chatbot can pull in relevant details from those past chats, so you don’t have to explain everything again.
Take a look at how the feature works in the video embedded above. Broadly speaking, Locked Chats keeps your private conversations out of that loop. The new privacy-centric feature is available now for every ChatGPT user, from the free tier to the Go, Plus, Pro, Business, Enterprise, and Education plans. If it’s not already available, try updating the app or wait a little bit, in case it’s a phased rollout.
OpenAI already offers a similar feature called Temporary Chats, which lets you have short-lived conversations with ChatGPT. You can choose whether a temporary chat is personalized or not. A non-personalized temporary chat ignores the AI memory, your custom instructions, and any apps you’ve connected to your ChatGPT account, so you start with a clean slate.