Algorithmic Bias in Hiring Algorithms: A Kenyan Perspective

Abstract

The use of machine learning algorithms has permeated into nearly all aspects of life. With this steady integration, tasks previously handled by humans are increasingly falling into the ‘hands’ of machines. Ideally this would be celebrated as a great improvement for mankind. Tasks that were previously riddled with human bias such as hiring would now be performed by an ‘omniscient algorithm’ that could harbor no bias therefore resulting in fair outcomes for the previously oppressed. However, this is not the case. The integration of machine learning algorithms in the hiring process risks further exacerbating existing bias that was prevalent or introducing new data-driven bias. The question then is how to contend with this novel form of discrimination: algorithmic discrimination. The answer to combating algorithmic discrimination is algorithmic fairness. The goal should not be to create ‘fair’ algorithms but rather to detect and mitigate fairness-related harms as much as possible. By doing so, a balance can be struck between the competing interests of innovation and employee rights. This article demonstrates that algorithmic discrimination during hiring is a real threat to the Kenyan jobseeker. Although this form of discrimination can be addressed by Kenyan law, more needs to be done to detect and mitigate fairness-related harms as much as possible.