We wanted to share an interesting recent Slate article entitled “Amazon Created a Hiring Tool Using A.I. It Immediately Started Discriminating Against Women,” by Jordan Weissmann (see https://slate.com/business/2018/10/amazon-artificial-intelligence-hiring-discrimination-women.html). The article highlights an important reality for employers, i.e., that algorithms used to automate certain processes, such as hiring, are not immune from discrimination.
Amazon implemented an automated system for reviewing and screening job applications for software engineers, but the system was programmed to look for terms found in applications for prior positions. The algorithm therefore perpetuated the male-dominated pool, essentially “learning” to discriminate against applicants whose resumes did not conform to that norm. Due to those issues, Amazon has scuttled the program.
As the article emphasizes, Amazon’s experience is an “illustration of why many tech experts are worried that, rather than remove human biases from important decisions, artificial intelligence will simply automate them,” linking to “Biased Algorithms Are Everywhere, and No One Seems to Care,” a July 2017 article in the MIT Technology Review by Will Knight (see https://www.technologyreview.com/s/608248/biased-algorithms-are-everywhere-and-no-one-seems-to-care/). In short, if you’re considering implementing algorithms for key personnel matters, be cognizant of the risks, including the continued risk of discriminatory decisions.