Antonym

Lab experiment

What can your AI see?

You’ve probably connected an AI assistant to your email, files or calendar. Pick what you’ve connected and see, in plain English, what it can read, what it can do, and what happens if someone tricks it.

What have you connected to an AI assistant or agent?

Everything happens in your browser. Nothing you pick is sent anywhere; we only count that someone used it.

Common questions

Is it safe to connect ChatGPT, Claude or Copilot to my email?
It depends on what else the assistant can do. Reading your email is one thing. The risk rises when the same assistant also reads content from strangers and has a way to send information out, because a stranger can then try to trick it into leaking your data.
What is prompt injection?
It is when text written by someone else, in an email, a web page or a shared document, contains instructions an AI assistant follows as if they came from you. The text can be hidden from people and still be read by the AI.
What is the lethal trifecta?
A term from security researcher Simon Willison for three things that are dangerous together: access to your private data, exposure to content from strangers, and a way to send data out. An assistant with all three can be tricked into leaking.