The Impacts of Algorithmic Work Assignment on Fairness Perceptions and Productivity: Evidence from Field Experiments
43 Pages Posted: 2 Apr 2020 Last revised: 12 Jan 2021
Date Written: March 8, 2020
With the increasing availability of data, the adoption of algorithms has become almost a necessity for businesses. Since algorithms often require human involvement, understanding how humans perceive algorithms is instrumental to the success of algorithm design in operations. In particular, the growing concern that algorithms may reproduce or even magnify inequality historically exhibited by humans calls for research about how people perceive the fairness of algorithmic decisions relative to alternative decision-making methods. We study how an algorithmic (vs. human-based) task assignment process changes task recipients' fairness perceptions and, subsequently, work productivity. We conducted a 15-day randomized field experiment with Alibaba Group in a warehouse where workers pick products based on orders known as "pickbills". Half of the workers were randomly assigned to receive their pickbills from a machine that ostensibly relied on an algorithm to distribute pickbills. The other half received pickbills from a human distributor. Despite using the same underlying rule to assign pickbills to workers in both groups, workers perceived the algorithmic assignment process as more fair than the human-based assignment process, causing a difference in perceived fairness by 0.94-1.02 standard deviations. This resulted in further productivity benefits: receiving tasks from an algorithm (relative to a human) significantly increased workers' picking efficiency by 17.3%-19.2%. The productivity gain from the algorithmic assignment was larger for more educated workers and workers who cared more about the difficulty of their pickbills, groups for which perceived fairness has a stronger effect on productivity. We replicated the main results in a second experiment.
Keywords: Behavioral Operations, Field Experiment, Productivity, Fairness, Artificial Intelligence
JEL Classification: C93, J24, O33
Suggested Citation: Suggested Citation