Adaptable (Not Adaptive) Automation: Forefront of Human–Automation Teaming

Abstract

Objective: Identify a critical research gap for the human factors community that has implications for successful human–automation teaming. Background: There are a variety of approaches for applying automation in systems. Flexible application of automation such that its level and/or type changes during system operations has been shown to enhance human–automation system performance. Method: This mini-review describes flexible automation in which the level of automated support varies across tasks during system operation, rather than remaining fixed. Two types distinguish the locus of authority to change automation level: adaptable automation (the human operator assigns how automation is applied) has been found to aid human’s situation awareness and provide more perceived control versus adaptive automation (the system assigns automation level) that may impose less workload and attentional demands by automatically adjusting levels in response to changes in one or more states of the human, task, environment, and so on. Results: In contrast to vast investments in adaptive automation approaches, limited research has been devoted to adaptable automation. Experiments directly comparing adaptable and adaptive automation are particularly scant. These few studies show that adaptable automation was not only preferred over adaptive automation, but it also resulted in improved task performance and, notably, less perceived workload. Conclusion: Systematic research examining adaptable automation is overdue, including hybrid approaches with adaptive automation. Specific recommendations for further research are provided. Application: Adaptable automation together with effective human-factored interface designs to establish working agreements are key to enabling human–automation teaming in future complex systems.