Carl Westin

2-2 Resolving automation acceptance issues 21 venting the inappropriate use of automation, 63 often categorized as misuse (overre- liance) or disuse (underreliance) of automation. 13 Research typically considers hu- man interaction with complex technologies capable of autonomy, in highly dynamic and complex environments characterized by high risk,. 13, 15, 64 Examples include ATC, aircraft carriers, nuclear power plants, space shuttle operations, fire fighting, and health care (see for example 65, 66 ). Several factors are believed to influence the choice of whether to use automa- tion. Examples are attitudes toward automation, trust, workload, complexity of automation, perceived risk of automation use, and perceived automation reliabil- ity. 13, 15, 16, 67–70 Riley argued that the core construct of automation reliance, defined as the “probability that an operator will use automation” [16, p. 21] is influenced by various factors such as trust in automation, self-confidence in manual performance, perceived risk, and fatigue. The framework suggested by Dzindolet et al. 55, 67 indicates that automation- use decisions, and which level of automation in particular, are determined by three decision-making processes (cognitive, social, and motivational) and their associated decision-making biases. At its core, the model proposes an evaluation of the per- ceived reliability of manual control against the perceived reliability of automated control. The outcome, measured in perceived utility of the aid, determines whether automation use is favored or not. This acceptance rationale, determined by a balanc- ing process weighing operator self-confidence against confidence in aid, is central in CE theories of automation use and trust. 63, 68, 70–73 Alternatively, reasons for automation resistance can already be viewed from a design perspective. Characteristics of poor compatibility might stem from the underlying goal for which the machine has been designed. That is, the deterministic algorithms embedded in automation generally aim to optimize. Such algorithms can be at odds with less structured, more heuristically governed human decision-making that tends to satisfice. Could there perhaps be an acceptance benefit if automation were designed consistent with human-like problem-solving styles? 2-2-1 Technology-centered automation Automation acceptance issues can be found in many different sociotechnical work domains in which skilled professionals are responsible for the safety and efficiency of operations. The ATC community has a well documented history of finding a suitable approach to automation design that promotes a functional and collaborative human-automation relationship. Over the years, innovative decision aiding systems and automation concepts have been proposed and developed to help controllers cope with the increasing pressures of the expanding ATC system (see 10 for an overview). Many current tools have grown from technology-centered research projects explor-