Fairness, Measured
Four numbers per group, and the first time two of them disagree
About 45 minutesMediumSociety and AI
What your child can do at the end
Compute selection rate, true positive rate, false positive rate and precision for each group
The brief
Eight hundred invented applicants, split into two invented groups, and a model that gives each of them a score. Nothing in this lab describes any real group of people: you choose how different the two groups' starting positions are, and you watch what that difference does to four measurements. Then you fix one of them and see what happens to the others.
Skills it develops
responsible aimodel evaluationstatistics basics
How it is assessed
3 graded challenges inside the lab, and 8 questions at the end that check the idea rather than the clicks. Every attempt is saved, so a retry after a mistake counts as the learning it is.