Amazon launched in 2014 a new algorithm for finding the best candidates. After a year, Amazon realized that the tool was biased towards women and shut down the program. John Jersin was the product manager for LinkedIn Talent Solutions when Reuters first broke the story in October last year. He offered his opinion on the current landscape of algorithmic hires: “I would certainly not trust any AI today to decide on their own.” “The technology just isn’t ready yet.”
In his comment, he implies that these systems will one day be ready. Adina Sterling is an assistant professor of organizational behavior at Stanford Graduate School of Business. Her research questions this optimism and links it to a deeply flawed misconception of the strategic role of hiring.
Sterling explains in a new article she co-authored with Daniel W. Elfenbein from Washington University in St. Louis and published in Strategy Science how smart hiring is intrinsically linked to long-term corporate strategies. She also explains why assigning the hiring responsibility to machines in the near term will likely compromise its strategic potential.
She says that “with technology increasingly playing this role, these questions of high-level strategy should be of great significance.”
Algorithmic Hiring: From Monster.com
Around a quarter of a century ago, career platforms such as Monster.com appeared online. These websites enabled companies to increase the speed and scale of recruitment. Once 20 people would apply for a job, suddenly 200 used in minutes. Sterling explains that “the vast majority of work has been devoted to this sourcing aspect” since then. It’s about filling the pool of applicants with the ones you need and separating the ones you don’t.
Between 65% and 70% are now handled by machines. In most cases, the best candidates are handed over to a person after an initial screening. While computers are becoming better at understanding data, and the amount of information available from our online footprints is increasing, the relatively strict filtering offered by many sourcing algorithms can be a problem. Sterling explains that it’s getting harder to find unique talent because these candidates don’t fall neatly into a single category. If someone with an improv background applied for a job in sales, a hiring manager would likely recognize their potential. A computer probably wouldn’t.
Artificial intelligence and algorithms are also slowly expanding from sourcing to selection. This practice is not yet widespread, but it will be soon. Sterling says that if hiring is to be considered strategic, leaving sourcing and selection to machines is a problem. Sterling says that humans still do much about strategy. Scholars need to consider what strategic managers are supposed to be doing when it comes to hiring.
Hiring as a Strategy
Sterling and Elfenbein define strategic hiring before claiming it. They claim that a decision is strategic if it meets four criteria.
Irreversibility: Irreversibility is the ability to make irreversible decisions. Once a decision has been made, all parties are bound to it. The choices of competitors are changed, and others are opened.
Interdependency means that a decision’s benefits depend on the relative alignment or complementarity between other decisions taken elsewhere in the organization. Interdependency puts a high value on sharing knowledge and coordination between decision-makers.
Competition: Competition is when competitors’ reactions influence the value of alternative alternatives. The decision-makers should, therefore, put themselves in their competitors’ shoes and predict how they would react to different options across various imagined futures.
Learning in an uncertain world: Employers consider the inherent uncertainty of hiring. They hire not only to perform a new task but also to learn new skills and connect with new people, communities, customers, and suppliers.
Sterling and Elfenbein interviewed hiring managers, AI experts, machine-learning experts, and executives from personnel-technology startups. They concluded that hiring should be considered strategic because it meets these criteria. “Things that can be programmed or automated — in other words, what machines already do in hiring — aren’t the strategic aspects of hire,” Sterling says that the strategic aspect of hiring includes addressing these four elements.
It is important to hire strategically, not only because of the skills of an individual applicant (the “best athlete” method) but also to find and integrate talent consistently and holistically. This means staying with the human capital strategy for a long, as these commitments cannot be retracted without cost.
Hiring is not just a search for the best candidate for a particular job. It’s a process similar to filling out a list. Hiring requires a global perspective of the company, its direction, and how it fits into a changing market. Computers lack this ability.
AI: What is its role
It is not meant to suggest that hiring algorithms wastes time. Sterling says that the role of hiring algorithms in any HR department must be matched with their strategic importance when deciding who a company employs. This means two different things.
Hiring managers should first take a close look at the algorithms that they use. She says that hiring managers might not be aware of the many filters an applicant pool may have gone through. While algorithms are efficient, many lose the nuance present in previous years when AI was less prevalent.
Managers should consider the value of their algorithm for hiring after determining if these filters need to be adjusted. Yes, it sorts resumes quickly. What else can it do? What does it NOT do? Then, they should integrate these insights into a broader view of the company’s strategic trajectory. How could a blind algorithm help or hinder the movement?
Sterling: “I’d love for managers to feel that they are held accountable for the actions of algorithms.” Sterling says that managers should at least understand this responsibility.
