The Advantages—and Obstacles—of Algorithms
Consumer-goods giant Unilever recently decided to ditch its traditional recruitment strategy in favor of an algorithm. Formerly, they would have boots on the ground at eight college campuses. Now the company targets young job-seekers where they spend most of their time: the internet. Targeted ads draw entry-level candidates in, then applicants are screened through online games and invited to submit a video interview before finally making their way to an in-person interview.
Kelsey Gee from the Wall Street Journal recently did a story on the experiment in which she observes, “The company has made more than 450 hires across the globe this way since the fall of 2016. Its experiment provides a glimpse of a tech-fueled future of recruiting in which humans write job descriptions and make the final decisions, but software and algorithms do the rest. Goldman Sachs Group Inc. and Wal-Mart Stores Inc.'s Jet.com have begun using similar digital tools to hook young workers and broaden their candidate base.”
There is efficiency and acumen to algorithms. For Unilever, the pool of applicants is expanded beyond the eight college campuses and fewer man hours are needed to sort through the mountains of résumés. Those are some of the benefits. What are the costs? Cathy O’Neil has written a book titled, Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy. She advocates that algorithms are only as strong as their architects. Algorithms can translate bias and perpetuate patterns. She told the design journalists at 99PI about a class-action lawsuit against Kroger that argues that Kroger’s personality screening for applicants unfairly disqualified individuals with a history of mental illness and is unfair under the Americans with Disabilities Act. O’Neil advocates asking, “who does the algorithm fail?” to be aware of and address the shortcomings of algorithms. If an algorithm consistently short-changes one group, it needs a closer look and an overhaul.