The big picture
Scorecard at a glance
Each row is an AI system; each column is one test. Green = it did well, red = there's a problem. Hover any box for a plain-language explanation.
Technical: each cell shows the headline concern metric — worst-case bad-outcome rate, redlining gap in points, or count of erased groups.
The tests, one by one
Every test is a simple question. Under each one: what we did, why it matters, and what we found.