Variable Selection in the Context of AI Fairness

Name: Variable Selection in the Context of AI Fairness
Uploaded: 2026-09-24T04:31:19.829682+00:00
Duration: 569 s
Description: The episode discusses a paper titled "Variable Selection in the Context of AI Fairness." The hosts argue against removing sensitive variables to achieve fairness, noting this can hurt model accuracy and allow bias to persist through proxies. They propose starting with all variables, optimizing for accuracy, and then using fairness audits to make targeted corrections.

The episode discusses a paper titled "Variable Selection in the Context of AI Fairness." The hosts argue against removing sensitive variables to achieve fairness, noting this can hurt model accuracy and allow bias to persist through proxies. They propose starting with all variables, optimizing for accuracy, and then using fairness audits to make targeted corrections.

Open the full episode