Hit 0.4% bias score on our hiring model, now leadership wants a party | Algorithmic Fairness Debate

Casey Thompson · 2026-09-01T23:50:14.327592+00:00

We finally got our hiring algorithm down to a 0.4% demographic bias score after 11 months of tweaking. That was way below the 2% target our legal team set, so I thought we'd just move on. But now my manager wants to put out a company-wide newsletter about it, and I'm worried that number doesn't tell the full story. We only tested it on 800 mock applications from our own applicant pool, not a real diverse sample. Has anyone else had leadership overcelebrate a metric that might not hold up in practice?

Free To Speak

Loading…

Hit 0.4% bias score on our hiring model, now leadership wants a party | Algorithmic Fairness Debate - Free To Speak