"Career coaches" are fear-farming the Stanford AI hiring study [debunk]
11 hours ago
- #Research critique
- #AI hiring bias
- #Algorithmic fairness
- The Stanford hiring study focuses on the pymetrics tool, not the entire AI hiring industry. It highlights flaws in its design where it learns from existing employees rather than predicting job performance.
- The study found that company-wide fairness audits can hide discrimination at the individual job level. When analyzing 1,746 positions separately, about 11% discriminated against Black applicants.
- Claims of 'algorithmic blackballing' or applicants being rejected everywhere are not supported by data: 84% applied to only one position, and a simulation showed no applicant was rejected by all models.
- The research is careful and includes limitations, such as lack of evidence on performance prediction and generalizability beyond pymetrics, but public commentary often exaggerates these findings.