Designing AI for All: A Primer on Bias in Artificial Intelligence Systems

Designing AI for All: A Primer on Bias in Artificial Intelligence Systems

By
Jennifer Aaker, Fei Fei Li, Thomas Higginbotham, Zoe Weinberg, Wendy De La Rosa
2020|Case No.ETH23| Length 4 pgs.

Algorithms improve our daily lives by filtering spam emails and curating our social media feeds, but they’re also used to recommend prison sentences, identify top job candidates, or guide autonomous vehicles. With each application, the risk of algorithmic bias varies from harmless inconvenience to life threatening and criminal. This case highlights key examples of unwanted bias in algorithms, outlines where the bias comes from, and presents several ways to begin combating this bias.

Learning Objective

This case provides an overview of artificial intelligence concepts that are critical for responsible and ethical leadership of the technology. It is meant as a primer with additional materials for deeper discussion and exploration into the topic.

This material is available for download at no charge.