Human-Automation Teaming: Unintended Consequences of Automation on User Performance

Abstract

The Federal Aviation Administration (FAA) Next Generation Air Transportation System (NextGen) is implementing new capabilities that rely on the support of automation to improve air traffic eficiency and predictability while maintaining safety. One issue that needs to be addressed regarding such new forms of automation - primarily as decision support tools - support is the possible impact on human performance such as knowledge and skill degradation. In particular, the question is whether after long-term use air traffic controllers will be able to effectively revert back to previously used manual operations if the automation becomes unavailable due to some degradation or failure. Based on this concern, this paper focuses on lessons learned from past experiences with the long term use of automation. To explore this issue, 21 experts were interviewed from four different domains: naval operations, medicine, nuclear power plant operation, and aviation. Three distinct job positions were interviewed within the domain of aviation: Airline dispatchers, pilots, and air traffic controllers. Two fundamental questions were addressed: 1.Can the introduction of automation have unintended negative impacts on human performance? 2.What mitigation strategies can be used to avoid or reduce these potential negative impacts on human performance and on system performance? Based on these data, five potential negative impacts of automation that were identified: •Degradation of Knowledge and Skill; •Reliance on Automation as the Primary “Agent” to Detect Problems; •Reduced Attentiveness and Preparation to Deal with Automation Degradations or Failures; •Alarm Fatigue, False Alarms, and Ignoring Alarms; •Automation-Induced Effects on Team Situational Awareness and Teamwork. In addition, six strategies to mitigate these potential negative impacts were documented: •Providing Effective Training and Experience; •Supporting Effective Teamwork; •Learning from Past Performance with Data Analytics; •Developing and Using Contingency Procedures and Checklists; •Designing Technology to Support Effective Human-Automation Teaming.