A Framework for Human-Algorithm Teaming in Biometric Identity Workflows

Abstract

Biometric identity workflows are made up of multiple subsystems which can be automated to varying degrees using computer algorithms. Setting the appropriate level of automation for each subsystem is crucial for optimal system performance, which relies on synergistic human-algorithm teams. In this work, we leverage an automation design framework from prior research to define levels of automation for each biometric subsystem. This framework aligns a four-stage model of human processing with equivalent system functions. We propose applying this framework as a method to determine the appropriate allocation of tasks between humans and algorithms within a biometric identity workflow. While previous work has focused on the role of the human in the comparison and decision subsystems, we emphasize the need to consider the full biometric system in determining the role of humans in biometric identity workflows.