Fuzzy

Fuzzy turns autonomous-car ethics into a compact series of forced choices. Each scenario asks which unavoidable harm a programmed vehicle should take, then uses the accumulated decisions to reflect a set of priorities back at you. It is more conversation starter than scientific instrument, and its omissions are often as revealing as its choices.

What is Fuzzy?

Fuzzy puts you behind a steering algorithm rather than a steering wheel. A collision cannot be avoided, so the task is to choose which direction the car should take and live with the principle hiding inside that click. One answer may protect passengers; another may spare more people; a third may reward those who followed the rules. Repeat the exercise and a rough portrait of your priorities begins to appear. The presentation turns an abstract debate into something immediate enough for a classroom, a group chat, or ten minutes of private discomfort. It also deserves skepticism. Player comments point out missing options and scenario details that can bend the apparent result, while the project publishes no visible methodology for its global-survey framing. Treat the outcome as a prompt for better questions, not a moral scorecard delivered from the dashboard.

What you can do there

• Choose between unavoidable outcomes
• Compare the values implied by choices
• Review a preference assessment
• Suggest new scenarios or improvements

Why we picked it

It makes a knotty policy question legible through small, concrete decisions. The useful part is not discovering a perfect answer; it is noticing when context changes your rule and arguing about why.

How to get the most from it

Answer quickly once, then repeat more slowly and name the principle behind each choice. Pay attention to scenarios where the offered paths feel incomplete; those objections are part of the exercise. Compare the final assessment with your actual reasoning rather than treating it as a verdict.

Good to know

The browser edition is an English Unity project first announced in 2018. A companion Android edition was updated in February 2026 and carries an Everyone 10+ mild-violence rating. Its store disclosure lists several collected data categories, but the browser version's current data handling and the survey's methodology were not established.

Who made it, and when?

Aurélio Santos is the credited creator or organization. The earliest supported launch date we found is November 13, 2018. A documented update was recorded February 23, 2026.