Help here, or there?

How The Charity Game works

You start with $1,000 in ten gifts of $100. At each stop, two people ask for the same $100 and you choose one. The other goes without, this time. There is no right answer, only yours.

Why it exists: decisions about aid money are usually made by a handful of people, or more and more by software. The Charity Game gathers what the public would choose, trade-offs included, from the players who choose to share.

Before you ride we ask for your gender (through your rider), age band, your region of the world and how often you give, each with a prefer not to say, so the results can be checked for balance. We never ask for your name, email, ethnicity, religion or health.

Each choice changes one or two things at a time: how many people are helped, what the money is for, whether it arrives as cash, goods or a service, whether you meet the person, their age, people or animals, proof that it works, and your own family against strangers. Every option is set in a real region of the world, so the dataset can show whether where people live changes who they help.

Sometimes details are hidden. You can look first or give without looking. Both tell us something.

Every ask is a campaign with a goal and a share already raised. The share is set at random, so we can see whether people back the causes that are winning or the ones left behind.

The kiosks you pass between stops show messages in a random order, and we note which ones you rode past before each choice. Rapid fire skips the ride and takes you straight from one choice to the next, which lets us compare the two.

How the questions are picked: first, four questions with the same number of people on both sides (your family, keeping the money, someone you met, your own region), so only that one thing is in play. Whatever you went with is then asked once more with four times the people on the other side. If you keep to it even then, that is one of your rules. Once taking more people has won three trade-offs, the game stops trading head counts and spends the rest on things it has not asked yet.

Your results show your rules first, then one bar per thing you were asked, with what you actually did written underneath. A fast answer counts as a clear preference and a slow one as a close call.

The Charity Game was inspired by MyGoodness, a game about how people give from the MIT Media Lab and The Life You Can Save. Sanctity is not affiliated with either. Everything in it is our own: how the questions are picked, the model that reads your choices, the things we test (real regions and prices, who is asking, proof, funding, family, your own region) and the results you see.

Prices by region are rounded from the World Bank’s price level ratios. The world map is drawn from Natural Earth (public domain). All other figures in this game are illustrative.

Your choices stay on this device. Nothing is sent anywhere.

A Sanctity AI Lab experiment, made by the team behind Sanctity AI.

Close