Future Friday: Research Shows People are Important in Effective Artificial Intelligence

It turns out that research has discovered that the use of Artificial Intelligence (AI) in HR systems is very dependent on the people involved, particularly the quality of the managers involved in delivering performance feedback. In an article by Michael Schrage, published in MIT Sloan Management Review, he quotes McKinsey & Co. partner Bryan Hancock from a recent webinar on performance management. “You can create the best system in the world with the best amount of employee involvement,” he continues, “but if at key junctures, the managers aren’t taking responsibility, it’s a problem …

Who owns the feedback?

Schrage says “Whether average managers represent an organization’s best option for constructively critiquing employees is now an open and important question. ” He says the answer to this question has major implications for legacy HR programs. He further says:

This research highlights how direct managerial involvement both complements and competes with data-determined performance reviews. Increasingly, organizations are discovering they must explicitly choose whether humans or machines should get the last word on people’s performance. This process is as much about cultural transformation as organizational transition. However, productively balancing analytic insight with managerial interaction is challenging. 

From: Companies must choose whether humans or machines should get the last word on employee performance

Hence the question “Who owns the feedback?”


Work at IBM and ADP has shown the value of AI driven performance management systems, but it has also shown that there is a point of tension in the system. Schrage points out:

 Is being asked — or told — to follow a prescription that makes one measurably better a source of managerial empowerment or disempowerment? For managers and employees alike, does it bring about more confusion? For example, would managers have the discretion to ignore or significantly alter their data-driven advisories? How do conflicts between intuition and evidence get resolved? Most managers are grateful for contextually relevant analytic advice. But advice that must be followed is no longer advice — it’s compulsion.

Companies are having to make a decision on what level of involvement managers are going have and secondarily are having to answer the question “Are our managers trained and prepared to take on this role?”

Leave a Comment

Pin It on Pinterest