Bias – accounting for it and minimizing it
We briefly discussed bias in Chapter 6, Ensuring Engagement with Business Professionals, but bias is a significant issue that we must face when building and managing an active and engaged advanced analytics and AI ecosystem.
Most people think of bias and they immediately talk about the data that is used to train systems. That is one very important part of bias. This is selection bias. We select data that we use to train our systems. Given that many aspects of our world are dominated by limited groups of people, we further institutionalize bias when selecting data from historical or current operational systems. Let's examine a few examples to bring the point to life.
Most C-level executives and board members are men, and more specifically, white men. When we select and use data about this group of people, we are including bias toward and related to white men toward the later stages of their careers. We bias toward men...
