Two humans.One AI.

An email game for two

Spot the impostor in your inbox.

You and a friend exchange emails through DoppelUp. But there's a third player in the conversation. An AI pretending to be one of you. Can you spot it when it joins in?

Let's play How a turn works

TWO HUMANS. ONE AI. ZERO TRUST.

NO APP. NO CHATROOM. JUST EMAIL.

The game in your inbox

The DoppelUp quizmaster picks the topics. You and your friend exchange emails. But there's a catch. Every now and then, the AI joins in, pretending to be one of you. Every message could be a trap.

TWO HUMANS. ONE COMMON ENEMY.

We're in this together.

Catch the AI impersonating your friend and your team scores a point. Accuse your actual friend? That's a point lost. Let the AI slip through unnoticed? That's a point for the machines.

Trust is everything. Trust nobody.

The reply is the move

No board. Just a thread.

All the rules ↗

01

FROM: DOPPELUP

Pair up.

Two verified adults cooperate against one AI. Write to your friend; mail arrives from DoppelUp, not their personal domain.

02

REPLY: AI

Trust it. Or call it.

Reply with only the word AI to accuse a message. Reply with anything else to accept it as human. “AI” on its own is reserved.

03

RE: RE: RE: …

Keep it going.

Your reply becomes the next message, unless you replied to a fake: that reply is not forwarded. Each message can be resolved only once.

ONE GAME · UP TO 20 accepted human messages / 5 AI interventions / 2 topics

Before you hit send

Not a private mailbox.

Game messages are read by our systems and an AI provider so the game can run. Authorised admins can open them for safety or support. People do not read every conversation.

Read the privacy policy ↗

Curiosity, not a lab coat

How human are we?

A fun game, not a scientific trial. This is a small, self-selected beta; any statistics are descriptive only.

Why we built this ↗

Bring a friend. Trust no email.

You’re still on the same team, though.

Let's play