Why It’s So Damn Hard to Make AI Fair and Unbiased

Author: Sigal Samuel

Publisher: Vox

Publication Year: 2022

Summary: The following article introduced the idea of bias in a statistical sense and how bias is interpreted in society. This is an important distinction to recognize and is often overlooked. The author discussed that defining what “fair” is can be very tricky and provided different definitions of “fairness”. An example is procedural fairness: an algorithm is considered fair if the procedure it used is fair, but an algorithm doesn’t know the underlying inequities in the procedure and does nothing to alleviate this.