why is my model unfAIr?
Published:
Algorithmic bias is an inherent challenge in machine learning models, particularly in high-stakes applications like credit approval, where fairness is critical. In this blog post, I discuss four key sources of bias in AI models: (i) bias in human decisions, (ii) proxy attributes, (iii) sample selection bias and (iv) algorithmic objectives. Each source of bias is accompanied by an empirical example from model-based credit approval.