Advancing technology has led to increased volume, variety, and velocity of the data being utilized in actuarial work. Because the actuary may be further from the collection of data than in the past, understanding what the data represents, its suitability, and its potential deficiencies— including bias—can be challenging. More generally, as more and more decisions in our lives are being decided by algorithms - from whether or not we’re qualified for a job to what clothes we buy to the medical treatment we receive, the importance of ensuring that these automated decisions are fair and ethical is reaching a greater urgency. In this presentation, panelists will discuss some of the key types of data bias that actuaries may encounter, and algorithmic bias will be dissected into main drivers with examples. The session will conclude with techniques for detecting algorithmic bias and best practices to remediate it
Learning Objectives:
Understand key types of data bias that actuaries may encounter and the underlying drivers of algorithmic bias.
Explain common metrics and their pros & cons to assess the presence of algorithmic bias.
Learn best practices for preventing algorithmic bias.
