Carragher and Hancock (2023) investigated how individuals performed in a one-to-one face matching task when assisted by an Automated Facial Recognition System (AFRS). Across five pre-registered experiments they found evidence of suboptimal aided performance, with AFRS-assisted individuals consistently failing to reach the level of performance the AFRS achieved alone. The current study reanalyses these data (Carragher and Hancock, 2023), to benchmark automation-aided performance against a series of statistical models of collaborative decision making, spanning a range of efficiency levels. Analyses using a Bayesian hierarchical signal detection model revealed that collaborative performance was highly inefficient, falling closest to the most suboptimal models of automation dependence tested. This pattern of results generalises previous reports of suboptimal human-automation interaction across a range of visual search, target detection, sensory discrimination, and numeric estimation decision-making tasks. The current study is the first to provide benchmarks of automation-aided performance in the one-to-one face matching task.

Identifying contacts in a military context can require operators to integrate multiple cues and to adjust response criteria to event base rates. The current experiment tested whether support from a decision aid would improve these processes. Participants performed a signal identification task that required them to integrate cues displayed as visual scale readings. In a static condition, participants saw a single set of readings each trial. In dynamic conditions, readings were updated over time. Base rates of signal categories were unequal, requiring participants to adopt biased response criteria to maximise response accuracy. Participants worked with or without an aid that combined cues and base rate information in an ideal manner. Support from the aid pushed participants\' response criteria towards optimal and improved integration of dynamic cues. Decision aids may be especially useful when task demands require biased response criteria and when cues are sampled over time.
Applied decision making often requires operators to gather and integrate multiple probabilistic cues. An experiment examined the information processing steps in multiple-cue decision tasks that could be improved by an automated decision aid. Statistically ideal aids improved operators’ response bias and information integration, although operator performance remained suboptimal.

UNASSIGNED: The exponential growth in computing power and the increasing digitization of information have substantially advanced the machine learning (ML) research field. However, ML algorithms are often considered \"black boxes,\" and this fosters distrust. In medical domains, in which mistakes can result in fatal outcomes, practitioners may be especially reluctant to trust ML algorithms.
UNASSIGNED: The aim of this study is to explore the effect of user-interface design features on intensivists\' trust in an ML-based clinical decision support system.
UNASSIGNED: A total of 47 physicians from critical care specialties were presented with 3 patient cases of bacteremia in the setting of an ML-based simulation system. Three conditions of the simulation were tested according to combinations of information relevancy and interactivity. Participants\' trust in the system was assessed by their agreement with the system\'s prediction and a postexperiment questionnaire. Linear regression models were applied to measure the effects.
UNASSIGNED: Participants\' agreement with the system\'s prediction did not differ according to the experimental conditions. However, in the postexperiment questionnaire, higher information relevancy ratings and interactivity ratings were associated with higher perceived trust in the system (P<.001 for both). The explicit visual presentation of the features of the ML algorithm on the user interface resulted in lower trust among the participants (P=.05).
UNASSIGNED: Information relevancy and interactivity features should be considered in the design of the user interface of ML-based clinical decision support systems to enhance intensivists\' trust. This study sheds light on the connection between information relevancy, interactivity, and trust in human-ML interaction, specifically in the intensive care unit environment.

By conducting a mixed-design experiment using simplified accident handling tasks performed by two-person teams, this study examined the effects of automation function and condition (before, during, and after malfunction) on human performance. Five different and non-overlapping functions related to human information processing model were considered and their malfunctions were set in a first-failure way. The results showed that while the automation malfunction impaired task performance, the performance degradation for information analysis was more severe than response planning. Contrary to other functions, the situation awareness for response planning and response implementation tended to increase during malfunctioning and decrease after. In addition, decreased task performance reduced trust in automation, and malfunctions in earlier stages of information processing resulted in lower trust. Suggestions provided for the design and training related to automation emphasise the importance of high-level cognitive support and the benefit of involving automation error handling in training.
The effects of automation function and malfunction on human performance are important for design and training. The experimental results in this study revealed the significance of high-level cognitive support. Also, introducing automation error handling in training can be helpful in improving situation awareness of the teams.

The concepts of automation trust and dependence have often been viewed as closely related and on occasion, have been conflated in the research community. Yet, trust is a cognitive attitude and dependence is a behavioural measure, so it is unsurprising that different factors can affect the two. Here, we review the literature on the correlation between trust and dependence. On average, this correlation across people was quite low, suggesting that people who are more trusting of automation do not necessarily depend upon it more. Separately, we examined experiments that explicitly manipulated the reliability of automation, finding that higher automation reliability increased trust ratings twice as fast as dependence behaviours. This review provides novel quantitative evidence that the two constructs are not strongly correlated. Implications of this work, including potential moderating variables, contexts where trust is still relevant, and considerations of trust measurement, are discussed.
Trust in automation is a cognitive attitude, and dependence on automation is a physical behaviour. Therefore, it is important to understand the differences between the two, especially as they have been conflated in the literature. This review highlights the small average correlation in the literature between subjective trust and objective dependence, which suggests that measuring trust as dependence (or vice versa) may not be valid. This suggests, then, that practitioners should carefully consider how trust and dependence are being measured in a given context so as not to incorrectly conflate the two.

Increased levels of digitalisation present major opportunities for efficiency in the oil and gas industry but can also contribute to new risks and vulnerabilities. Based on developments in the industry, the Norwegian Ocean Industry Authority (HAVTIL) has in recent years pursued targeted knowledge development and follow-up of company\'s digitalisation initiatives. This paper explores data collected through HAVTIL\'s audits of the development and use of automated systems within well operations. The analysis of the data resulted in the identification of five main topics related to the implementation of digital technologies. The five main topics were organisational complexity, follow-up and implementation of technology, analysis and documentation, user-interface and alarms and competence and training. Overall, the results support research findings within human factors and technology development, pointing out that there is a lack of focus on human factors in both development projects and in operations. In addition, this paper provides insight into how digitalisation initiatives are followed-up and explores the results from the analysis in light of the current developments in the industry.
To investigate automated operations and human performance, three audits were performed by the Norwegian Ocean Industry Authority (HAVTIL). These audits have been used as case studies and the basis for this paper. Results from the analysis support research findings within the field of human factors and technology development, pointing out that there is a lack of focus on human factors in both development projects and in operations.

OBJECTIVE: The primary purpose was to determine how trust changes over time when automation reliability increases or decreases. A secondary purpose was to determine how task-specific self-confidence is associated with trust and reliability level.
BACKGROUND: Both overtrust and undertrust can be detrimental to system performance; therefore, the temporal dynamics of trust with changing reliability level need to be explored.
METHODS: Two experiments used a dominant-color identification task, where automation provided a recommendation to users, with the reliability of the recommendation changing over 300 trials. In Experiment 1, two groups of participants interacted with the system: one group started with a 50% reliable system which increased to 100%, while the other used a system that decreased from 100% to 50%. Experiment 2 included a group where automation reliability increased from 70% to 100%.
RESULTS: Trust was initially high in the decreasing group and then declined as reliability level decreased; however, trust also declined in the 50% increasing reliability group. Furthermore, when user self-confidence increased, automation reliability had a greater influence on trust. In Experiment 2, the 70% increasing reliability group showed increased trust in the system.
CONCLUSIONS: Trust does not always track the reliability of automated systems; in particular, it is difficult for trust to recover once the user has interacted with a low reliability system.
CONCLUSIONS: This study provides initial evidence into the dynamics of trust for automation that gets better over time suggesting that users should only start interacting with automation when it is sufficiently reliable.

Vehicle automation is becoming more prevalent. Understanding how drivers use this technology and its safety implications is crucial. In a 6-8 week naturalistic study, we leveraged a hybrid naturalistic driving research design to evaluate driver behavior with Level 2 vehicle automation, incorporating unique naturalistic and experimental control conditions. Our investigation covered four main areas: automation usage, system warnings, driving demand, and driver arousal, as well as secondary task engagement. While on the interstate, drivers were advised to engage Level 2 automation whenever they deemed it safe, and they complied by using it over 70% of the time. Interestingly, the frequency of system warnings increased with prolonged use, suggesting an evolving relationship between drivers and the automation features. Our data also revealed that drivers were discerning in their use of automation, opting for manual control under high driving demand conditions. Contrary to common safety concerns, our data indicated no significant rise in driver fatigue or fidgeting when using automation, compared to a control condition. Additionally, observed patterns of engagement in secondary tasks like radio listening and text messaging challenge existing assumptions about automation leading to dangerous driver distraction. Overall, our findings provide new insights into the conditions under which drivers opt to use automation and reveal a nuanced behavioral profile that emerges when automation is in use.

Real-world events like the COVID-19 pandemic and wildfires in Australia, Europe, and America remind us that the demands of complex operational settings are met by multiple, distributed teams interwoven with a large array of artefacts and networked technologies, including automation. Yet, current models of human-automation interaction, including those intended for human-machine teaming or collaboration, tend to be dyadic in nature, assuming individual humans interacting with individual machines. Given the opportunities and challenges of emerging artificial intelligence (AI) technologies, and the growing interest of many organisations in utilising these technologies in complex operations, we suggest turning to contemporary perspectives of sociotechnical systems for a way forward. We show how ideas of distributed cognition, joint cognitive systems, and self-organisation lead to specific concepts for designing human-AI systems, and propose that design frameworks informed by contemporary views of complex work performance are needed. We discuss cognitive work analysis as an example.
Emerging developments in AI will pose challenges for the design of human-machine systems. Contemporary perspectives of sociotechnical systems, namely distributed cognition, joint cognitive systems, and self-organisation, have design implications that are unaccommodated by traditional methods. Cognitive work analysis may provide a way forward.Abbreviation: AI: Artificial intelligence.

The aim of the current study is to investigate predictors and consequences of driver-initiated take-overs during automated evasion maneuvers. Literature on control transitions in automated driving has mainly focused on system-initiated take-overs. However, drivers may also initiate take-overs without take-over requests. To date, such driver-initiated take-overs have rarely been investigated. Our study addresses this research gap. In a driving simulator study with 61 participants, we investigated whether the criticality of highly dynamic evasion maneuvers and trust in automation affect the probability of driver-initiated take-overs. Criticality was manipulated via time headway (THW) and traction usage (TU). Trust was varied by manipulating automation reliability before the experimental trials. Consequences of driver-initiated take-overs in terms of collisions and lane departures were assessed. The results indicate that THW, TU, and trust affect the probability of driver-initiated take-overs. Moreover, the time it takes the automation to respond to an obstacle ahead by starting an evasion maneuver may be another relevant factor in predicting take-overs. After a take-over, drivers produced a number of unnecessary lane departures and collisions. These were independent of THW and TU. The study demonstrates that drivers are more likely to take over vehicle control during automated evasion maneuvers without take-over requests when criticality increases and trust in automation decreases. Such take-overs may be hazardous for traffic safety. Our findings help to design automated vehicles that avoid unnecessary take-overs in critical driving situations or de-escalate their consequences effectively, thus increasing traffic safety.