Bias and Fairness of AI-Based Systems Within Financial Crime

Author: Danny Butvinik

Publisher: NICE Actimize

Publication Year: 2022

Summary: The following article discusses the benefits of addressing bias in artificial intelligence (AI)-based systems and provided a “Fairness-Aware Culture” guideline for addressing this issue. The guideline focused on both the level of nontechnical self-assessment and the level of technical controls and means of evaluation. The guideline consisted of 5 elements of data fairness that data science teams should keep in mind: 1). Correct representational flaws in the sampling; 2). Determine fit-for-purpose and sufficiency of data quantity; 3). Securing source integrity and measurement accuracy; 4). Scrutinize the timeliness and recency of the dataset; and 5). Need domain expertise to assist in modeling and sources of measurements.


Posted

in

, , ,

by