I compared the privacy of ChatGPT, Gemini, Claude and Perplexity — here’s the one you should trust most with your personal info

If you use AI chatbots for anything remotely sensitive — finances, health questions, job concerns, parenting, relationships — choosing the right one matters more than you think. Every LLM handles your data differently. Some save your chats forever. Some send them to human reviewers. One doesn’t train on your data at all.

I dug through the privacy practices of the four biggest AI chatbots — OpenAI’s ChatGPT, Google’s Gemini, Anthropic’s Claude and Perplexity — to see what really happens to your conversations behind the scenes. The differences were bigger than I expected.

Here’s how each chatbot treats your data and which one is the safest choice if privacy is your top concern.

ChatGPT: the least private AI

ChatGPT is the most widely used chatbot — and also the least private unless you actively change your settings. By default, ChatGPT uses your conversations to train its models. Your chat history is saved indefinitely unless you manually delete it.

OpenAI does offer Temporary Chat, which prevents your conversation from being saved or used for training. But you have to remember to turn it on each time you need it.

To OpenAI’s credit, the controls are there, but privacy requires deliberate effort. For everyday use that isn’t sensitive, it’s great. For private matters, you may want to think twice.

Why ChatGPT ranks last