This article dives deeper into algorithmic bias which is bias that artificial intelligence programs adopt to reflect societal patterns. The paper argues that algorithmic bias is harder to identify due to it correlating socially-senstive attributes to one and other. This, however, makes it hard to fix due to the algorithm needing these attributes. The paper believes this problem is similar to the ones in the cognitive domain as through analysis their comparisons are positive. The paper, however, clarifies that the paper is not intended to argue that algorithmic and cognitive biases are similar in all fronts but merely similar on some fronts. They continue to argue that acknowledging this similarity can help us understand the nature of bias better.
