Biases in AI-Based Hiring Algorithms: Effects on Minorities and How It Can Be Eradicated

Abstract

In today’s society, many companies are utilizing artificial intelligence to hire people, making this process much easier and more efficient. This has deeply changed and impacted the way that companies look for potential employees. However, the hiring algorithms have received extremely biased datasets containing inaccurate or incomplete information,  putting minorities at a disadvantage because their application reveals their gender and race. This has cost many applicants opportunities, while also giving opportunities to people who may not be qualified. To understand the biases much more deeply, this paper discusses examples of the types of biases that can be seen, multiple AI-based hiring technologies developed, and the advantages and disadvantages of them. It also discusses a study utilizing both quantitative and qualitative aspects that was conducted. Surveys were sent out to employees of the technology industry to gain feedback about the algorithms, while data showing the percentage of ethnic groups, females, and males hired from multiple companies was analyzed. The companies are compared and contrasted with each other to show better insights on which hiring algorithm for each company is more effective. It is found that there is not a significant difference in the profile of the technology industry. However, in some areas, there was data that did not account for hiring patterns in certain ethnic groups due to biased datasets.  Ultimately, a plan is devised to enhance current hiring algorithms and potentially develop a new one to promote inclusivity and fairness.