AI Adoption Can Backfire: How Leveling the Playing Field Can Demotivate Top Employees

Companies are racing to roll out new artificial intelligence technology to replicate what star employees do to boost the bottom line. But that can backfire and actually hurt the firm more than benefit it, finds new research from the University of Maryland’s Robert H. Smith School of Business. In competitive workplaces, managers need to be very careful about how they implement and tune AI to keep the best performers motivated and keep overall productivity from declining.

Assistant professor Manmohan Aseri, co-author of the research forthcoming in Management Science, looked at environments where employee compensation is strongly tied to performance goals, such as in sales (although results are applicable in many general settings). Those firms may push top employees to use and train AI tools. When other employees at the firm use the AI—that now includes all the star employee’s tricks and tactics—the playing field gets levelled. And in firms with competitive pay-for-performance arrangements, employees are often competing with each other for bonuses. 

“Those star employees basically lose their competitive advantage,” Aseri says. “And because of that, they might lower their effort.”

If the employees training the AI stop doing a good job, the AI will not be good anymore either, Aseri says. If it doesn’t give good recommendations, the other employees using it suffer, and the firm’s overall productivity goes down.

“The benefit coming out of AI adoption may be less than the harm coming out of these employees being demotivated because of their loss of their competitive advantage,” he says.

To keep that from happening, Aseri says organizations have to do more than just pay the high performers a guaranteed salary—because that doesn’t work. “The moment you guarantee anything, it reduces the competition among employees and motivation goes down anyway.”

Aseri says the only strategy to effectively mitigate this issue is to deliberately “dumb down” the AI—adopting it only at a partial scale.

That means leveraging AI to help, but not actually enough to level the playing field.

“Use the AI to help a person who is not doing very well, but don’t help so much that they become better than the star employee,” Aseri says.

For employers, the key message is to slow down the AI rush, says Aseri. He calls for a “graded adoption of AI,” rather than a “knee-jerk” overnight rollout, which is currently the trend in many industries. He recommends that managers take top performers into their confidence to explain the goals of AI implementation while assuring them of their job security and their value to the organization. He says firms should consider a targeted approach, where the AI benefits are only for the people who trained it, at least initially.

For top employees, Aseri says the research offers a powerful negotiation tool: Demand protection of exclusivity over the AI trained on their knowledge and skills to make sure they stay valuable to the organization—and motivated to continue to perform. Afterall, it’s their human expertise that adds up to the success of any AI implementation. 

The study also highlights the changing nature of essential skills to be a top performer in the age of AI. 

“In the pre-AI era, highly technical skills were more rare and very valuable,” Aseri says. 

Now that AI excels at technical skills, the “soft skills”—empathy, reading a customer, knowing “when to back off, when to push a bit more,” and the nuances of human interaction—are more important than ever, says Aseri.

“The human skills—the interpersonal skills—are going to be the true competitive edge.”

Read Aseri’s research, “Backfiring AI? AI Deployment in Workplace,” co-authored with Di Yuan of Auburn University and Narayan Ramasubbu of the University of Pittsburgh, forthcoming in Management Science.