Methods to help workers perform better commonly focus on how the group improves on average, neglecting to account for how different types of employees might benefit.
A management researcher at The University of Alabama is part of a federally funded project to use machine learning to examine how improvement methods affect all workers.
“We’re looking at the highest and lowest performers to see if their performance is actually improved by interventions,” said Dr. Kimberly Stowers, a UA assistant professor in the Culverhouse College of Business. “We want to overcome the tyranny of averages to account for people on the edges, so no one is left out.”
Stowers is part of a program funded by the Defense Advanced Research Projects Agency, commonly called DARPA, part of the federal Department of Defense. The 18-month project will use teams across the country to develop machine learning that considers all facets of improving human performance.
The military’s interest is to optimize human performance by understanding which intervention methods can be most effective for different individuals and teams, but the implications of the research would help all organizations, Stowers said.
The work will use, as DARPA calls it, “third wave” artificial intelligence where computer programs adapt to situations through contextual reasoning that could not come from simply analyzing large data sets to reach a conclusion.
“It’s cutting edge, which makes it risky,” Stowers said. “There are different ways of improving human performance that we know we can model, but, in this case, we don’t know if we can model the methods effectively.”
Stowers and a graduate student researcher, Lisa Brady, will inform the AI tools by providing a theoretical framework of individual differences that shape how people work. They will also use information provided by people who have studied team performance to enhance the algorithms used in the project.
Methods to improve performance include training, feedback, practicing a skill and a host of other strategies designed to help workers do their jobs better. The questions are how these strategies improve performance for all workers and whether AI tools can predict and take advantage of the differences among workers in their response to improvement strategies.
This information will be provided to an artificial intelligence company that will develop the “third wave” AI algorithms. The goal is to have a computer program that can inform workplace managers about the likelihood of a strategy working for the individual employees, Stowers said.