Carl Westin

86 Automation transparency effects dynamic, safety-critical and high-risk control room environments. One such do- main is ATC, currently facing a large and necessary technological modernization expected to considerably change how controllers interact and communicate with automation. 3, 4 Achieving an effective and functional team relationship between the controller and automation has proven difficult. Much advanced automation intro- duced to current ATC have been distrusted and even rejected. 14, 189 Practically, it is possible to address this issue by considering the degree of au- tomation transparency afforded by the interface. 190–193 Automation transparency is generally concerned with the system’s ability to afford understanding about how it works, its behavior, and its intentions. It is typically achieved by provid- ing explanations for system reasoning or visualizing automation meta-information in relation to, for example, information uncertainties and automation reliabil- ity. 191 Empirical research indicates that more transparent systems can benefit human-automation cooperation in terms of, for example, performance, 191 accep- tance, 111, 194, 195 trust, 174, 191, 195 and understanding of advisories. 111, 196 As an alternative to automation transparency, increasing interest has been given to personalizing automation, and in particular decision support systems, to individual preferences, needs, and abilities in task performance and problem- solving. 154, 155, 197, 198 Empirical studies we conducted in the context of ATC CD&R, suggest that strategic conformal resolution advisories (i.e., personalized) benefit controllers’ acceptance and agreement with a decision aid, while also reducing their response time to advisories. 101 Such conformal advice can be considered transpar- ent in that it may be perceived and interpreted by the controller to afford a rationale underlying the system’s reasoning, simply because it matches with how the con- troller would solve the conflict. At a general level, this paper focuses on two complementary ways for improving the operator-automation relationship: by exploring individual differences in person- alized decision support, and interface transparency of meta-information visualiza- tion that facilitates operator understanding of the system. More specifically, this study investigates how controller performance in a CD&R task, together with the acceptance and interaction with a decision aid, is affected by interface transparency and the strategic conformance of the system’s resolution advisories. It is hypoth- esized that a more transparent interface will reduce the importance of conformal decision support, be more readily understood and positively affect acceptance of performance in a conflict resolution task. While ATC is the focus of this paper, the results are applicable to several other domains with similar dynamics, task demands, and reliance on highly skilled operators.