Individual Submission Summary
Share...

Direct link:

Preventing Institutional Power Abuse with AI in the COVID-19 Pandemic

Thu, October 7, 1:20 to 2:50pm EDT (1:20 to 2:50pm EDT), 4S 2021 Virtual, 15

Abstract

Distributing shifts evenly without the help of an AI is a challenging process. Furthermore, the process’ difficulty grows exponentially with the increase in number of personnel involved and externalities such as the harsh working conditions of COVID-19 pandemic. Access to such sophisticated software is nearly impossible therefore, neither the preparation nor the crosschecking of shift schedules are handled efficiently due to the time limitations in health institutions. Thus, without any digital solution present, the distribution process becomes open to human bias and exploitation. To address this problem, we have developed a semi-automated workforce scheduling software which guarantees distributive justice with the help of mathematical models and machine learning algorithms. By digitally transforming the biased methods of schedule preparations of three sample hospitals, this study aims to critically analyze possible shifts in power relations of organizational hierarchies and suggest methods for preventing power abuse in organizational medium with the help of sociological theories such as ‘structuration theory’ and ‘social dominance theory’. According to our preliminary findings, many of the responsible health personnel use this gap in supervision by favoring their equally ranked (or aged) peers by prioritizing them with more preferred leave days and assigning them more frequently with paid extra shifts. This inevitably causes decrease in morale, worker fatigue, and tensions among health professionals. The research in biases in algorithms is a hot topic; however, in this study we adopt the opposite approach and employ AI in order to reduce human error, bias, and exploitation.

Authors