Human Resource (HR) managers often view their job as dealing with people. They believe that in order to do it well, they must know people, read emotional cues, and respond with empathy. Therefore, they conclude, should artificial intelligence (AI) put jobs at risk, they’ll be the ones delivering bad news, not receiving it.
Well, I am sorry to be the bearer of bad news, but HR jobs are actually quite likely to be impacted by artificial intelligence.
Often our view of AI comes from the popular culture. Whether smart machines be threats (like the Terminator or Ultron) or compatriots (like Star Trek’s Data or Lost in Space’s Robot), they are most often characterized by emotion detachment or a struggle to interpret human emotions. In other words, no one is giving these robots a job in HR.
But the recent developments in AI have not been about creating the type that simply automate what a human can do. Instead, what current advances in machine learning and neural networks do is make prediction better, faster, and cheaper.
Broadly speaking, prediction is when you take information you have (e.g., data on past weather or classifications of photos) and turn it into information you do not have (e.g., what the weather will be next week or what is in this particular image).
Making good predictions is the core of a good HR manager’s job. HR managers have to predict whether a candidate’s CV makes them worth interviewing. They have to predict whether, based on that interview, a candidate may be appropriate for a job. They have to predict whether a performance evaluation indicates that an employee should get a pay rise or a promotion. And they have to predict whether an employee’s concerns about his or her boss are legitimate or not.
While a job that involves hiring people might feel as though it requires human intuition, objective statistics prove more effective. In a study of hiring across fifteen low-skilled service firms, my colleague Mitch Hoffman along with Lisa Kahn and Danielle Li found that firms using an objective and verifiable test along with normal interviews, gained a 15% jump in the job tenures of hired candidates relative to just using interviews alone. This is despite maximizing tenure being the main goal of hirers.
Prediction machines feed off data, and in the HR space, the data is available. Based on it, increasingly complex algorithms will be generated to help HR departments with their predictions. In so doing, these could reduce bias, errors, and time in the evaluation of people. In other words, these algorithms have the potential to be better and quicker than people in those tasks.
This means that AI will almost certainly impact HR jobs. But there’s good news, too: Our experience elsewhere suggests that as jobs transform to accommodate new technology this exposes the real human element behind those jobs. It may well be that a human face is still needed to deliver the news even if it is a machine generating the news.
Joshua Gans is professor of strategic management at the Rotman School of Management, University of Toronto and the author of Prediction Machines: The Simple Economics of Artificial Intelligence.