Ready or Not, Here Comes Algorithmic HR

The July issue of The Atlantic features a really interesting article that foreshadows a major shift in the practice of human resources: the rise of the algorithm and the fall (or at least dramatic reduction) of reliance on human judgment in hiring, assessment, and potentially a host of other HR applications.

The title of the article, “For More Workplace Diversity, Should Algorithms Make Hiring Decisions?” turns out to be a rhetorical question, because the author’s view is…absolutely yes. The examples cited make a compelling case for the value of algorithms in increasing the diversity of hiring. And of course it’s a very short leap from using algorithms for hiring to using them for assessment and talent development. They already are present (because algorithms are, after all, just complicated sets of rules) in compensation and benefits and performance management.

So is that the end of the story? Should we just welcome the brave new world of Algorithmic HR?

Well yes and no. Yes, because algorithms are a powerful way to compensate for well-recognized weaknesses in human reasoning. These include the unconscious biases that contribute to reducing the diversity of candidate pools, and the confirmation bias – the tendency to gather data that confirms our judgments.

In general, our brains are not wonderful at dealing with complexity and it’s here that algorithms shine. So their analyses can incorporate many more variables than our minds. And algorithms don’t have biases unless they are explicitly built into the rules.

However there are some questions that need to be asked before we rush headlong into the embrace of the algorithm. Questions such as:

Are algorithms easier to “game” than people? Because the first thing we should expect after people learn that hiring or appraisal or development is governed by a set of rules is efforts to understand what those rules are and tailor inputs to improve attractiveness. How do we avoid an “algorithmic arms race“ with people developing algorithms that try to game the other algorithms?

What do algorithms do with the outliers? The goal in building algorithms typically is to increase average performance. This is of course a good thing, especially when it eliminates the below average. But will this come at the expense of investing in people who have the potential to be exceptional? Will reliance on algorithms mean companies stop making high-risk, high-return bets on people?

What happens when everyone uses algorithms? The examples in the Atlantic article showcase the benefits of algorithms for increasing hiring diversity at the organizations that use them. But they work well given that most other organizations are not using them. What happens when everyone is using the same algorithms to seek the same sorts of people? The results in the stock market are instructive. Early adopters of algorithms made huge returns…and then everyone else caught up.

What do you think of the future of algorithmic HR? Do you have other examples? Do you see other benefits or have other concerns?

Dr. Michael D. Watkins is Professor at IMD business school and the co-founder of Genesis Advisers and the author of The First 90 Days, Your Next Move and numerous other books and articles about leadership transitions and onboarding. Her moderates The First 90 Days group on LinkedIn #First90 #onboarding #HR