Author: Terence Shin
Publisher: Medium
Publication Year: 2020
Summary: The following article begins by giving a brief overview of how dangerous artificial intelligence (AI) models can be if they are not properly checked for biases, then it dives into 3 real-world examples of actual ethical violations and biases that have happened in decision-making processes in real companies. The first involved a healthcare service favoring patients for needing extra care unfairly on the basis of race. The second involved an algorithm that predicted if a detained criminal was likely to commit another offense – again resulting in an unfair false positive discrepancy on the basis of race. Lastly, Amazon was mentioned as having a bias in their hiring process that created an unlevel playing field for women. AI algorithms can create extremely serious consequences for real people living out their daily lives. It is essential to monitor all algorithms and models for bias before allowing them to make such impactful decisions for the public.
