The dynamics of trust: Comparing humans to automation.

Abstract

Complex industrial environments involve cooperation between operators and automation. The strategies used to allocate tasks to automation are a crucial component of that cooperation and are known to be affected by the operators' trust in the automation. In 2 simulated process control experiments, the authors compared trust in automation with trust in human partners in equivalent situations. Experiment 1 found the relationship between trust and task allocation to be qualitatively identical, but quantitatively attenuated, for human partners as compared with automation. Experiment 2 additionally identified the operators' trustworthiness, as they thought it would be perceived by a human partner, as crucial to task allocation under human collaboration but not under automation. The results imply that human collaboration benefits from calibration of people's assessment of how others perceive them.

Making adaptive cruise control (ACC) limits visible

“…However, drivers also relied on the ACC more during the first half of the rain/ traffic periods than during the second half. This nonlinear reliance across time shows that reliance does not recover as quickly as reliability, a finding consistent with that of other researchers (Lee and Moray, 1994;Lewandowsky et al, 2000;Lee and See, 2004). Reliance results for failure and non-failure periods together provide a high-level representation of how often drivers use the ACC and in what operating condition: normal or failure.…”

Drivers’ Reliance on Lane-Keeping Assistance Systems as a Function of the Level of Assistance

“…On the contrary, if operators are more confident in their own abilities than they trust the automation, then they will likely choose manual control. Empirical evidence for this was found in a number of studies (e.g., De Vries, Midden, & Bouwhuis, 2003;Lee & Moray, 1992, 1994Lewandowsky, Mundy, & Tan, 2000). Riley's model also accounts -although not very detailed -for the influence of operator states (fatigue) on reliance and the time dynamics influencing the relation between trust and reliance.…”

“…Whereas Muir and Moray (1996) found that operators' trust was a very good predictor of their reliance, others (De Vries et al, 2003;Lee & Moray, 1992, 1994Lewandowsky et al, 2000) found that operators' reliance could be more appropriately predicted by the relationship between their trust in the automation and their self-confidence in manually controlling the system. Specifically, Lee and Moray (1994) found that participants' reliance could be predicted by a logit function of the difference between trust and self-confidence, whereby in general operators tend to rely more on automation when their trust in the automation exceeds their self-confidence, and tend to engage in manual control when trust was lower than their self-confidence.…”

“…Substantial empirical evidence was found that trust in an automated system varies as a function of its reliability (De Vries et al, 2003;Lee & Moray, 1992, 1994Lewandowsky et al, 2000;Metzger & Parasuraman, 2005;Moray, Inagaki, & Itoh, 2000;Muir & Moray, 1996). In many of these studies participants acted as supervisors of a simulated process control plant in which different subsystems could be operated under automatic or manual control.…”

Trust in vehicle technology

“…The social-psychological component is captured in the intended behaviour being situated within a set of socially defined norms governing whether a behaviour is normal/acceptable (Lewandowsky, Mundy & Tan, 2000). Finally, and importantly for trust, is the emotional/affective component captured under the various attitudes the driver has about the behaviour (Merritt et al, 2013).…”