Incognito AI Chat. Is It Really Private?

A lot of people think incognito mode means the AI chat just disappears. Not quite.

Both of the big chat apps have a private mode now. Claude calls it incognito, ChatGPT calls it Temporary Chat, and I use one of them almost every day. They do real things. But "private" here means something narrower than most people assume, and it's worth knowing where the line is before you paste something sensitive into one.

Where to find it

In Claude, start a new chat outside of a project and look for the ghost icon in the top right corner. Click it and the screen tells you you're incognito.

In ChatGPT, open a new chat and click the Temporary button in the top right. The page switches to Temporary chat.

What it actually does

In both apps the chat stays out of your history, it doesn't become a memory the assistant carries into later conversations, and it isn't used to train the model.

One wrinkle on the ChatGPT side. If your personalization is on, a temporary chat can still read the memories and custom instructions you already have. It just won't save anything new from that conversation.

The catch

A copy can stick around.

OpenAI says ChatGPT may keep a copy of a temporary chat for up to 30 days for safety. Claude holds incognito chats for up to 30 days too. And if a Claude chat gets flagged by its safety systems for a policy violation, it can be kept for up to two years.

Then there's work. If you're on a company account, your company's retention rules still apply. The toggle on your screen doesn't override what your employer has set.

Why I use it

I use ChatGPT a lot, and I go back to old chats all the time. Incognito keeps that list clean. The one-off questions and random "how do I" stuff stop piling up in the sidebar between the conversations I actually want to find again.

The trade is you can never reopen it. Once you close an incognito chat in Claude, it's gone from your side. ChatGPT works the same way unless you save the temporary chat before you close it.