Article Summary

VISIT AI Trust Library
Article Summary of:

RDS and Trust Aware Process Mining: Keys to Trustworthy AI?

Published:
January 24, 2022

Artificial intelligence is a highly innovative technology that helps businesses expedite processes and increase efficiency. In turn, companies are predicted to invest $500 billion annually in AI by 2024. However, AI introduces unique concerns into business models. Most concerning as of late is the ways in which AI amplifies pre-existing biases on a large scale. Though trust aware processes that integrate visualization, discovery, analysis, human-centric examination, and monitoring help to mitigate risks associated with AI bias, enterprises must understand what AI fairness is to create a comprehensive plan to institute it. 

In his latest article in Techopedia, Andrew Pery outlines the five challenges associated with applying fairness to AI systems:

  1. The concept of “fair” is interpretable based upon cultural, social, economic, and legal boundaries.
  2. Fairness and bias are not always the same idea. 
  3. To provide group fairness or individual fairness, businesses must approach their AI strategies differently.
  4. To ensure fair outcomes, statistical parity must be balanced.
  5. Fairness is defined by those in power, which can perpetuate pre-existing power hierarchies.
Text Link
Ethics & Responsibility