Enough Fluff: Returning to Meaningful Perspectives on Automation

As society progresses towards increased automation in aviation—such as with Advanced Air Mobility and Unmanned Aircraft Systems—it is important to have a common understanding and perspective about automation among the many stakeholders, including aviation system designers, operators, maintainers, and regulatory authorities. Unfortunately, the discourse is hindered by misleading perspectives, assumptions, claims, and terminology.There are many examples. The term "automation" can be simply defined, but it is often confounded with "autonomous" and other descriptions of the function being automated, and further confounded by our subjective opinions on which functions are considered "advanced" or "intelligent". Automation is often discussed not as a tool that can be leveraged to achieve goals of the aviation community, but rather as a technocentric goal in itself. We often refer to automation as "an AI" (artificial intelligence) or a "team member", or other ways in which we anthropomorphize machines, yet do not clearly define functions for automated components of these desired systems. We argue that humans are prone to errors and that more automation therefore means fewer errors, without a fair balance that considers humans as valuable functional elements. We talk about operator trust as if the idea is unique to AI, when in fact the basic principles for human-automation interaction have not changed. We try to treat automation as a one-dimensional variable, such as with automation levels, but this hides important detail and has limited value in applications such as design, operations, and approvals of complex human-automation systems.This paper identifies issues in recent automation and human-automation discourse, and provides clarifications and recommendations to improve progress towards the integration of increased automation in aviation systems.