Accumulating Distrust: A Dynamic Model of Automation Disuse

Abstract

Human-automation driving style similarity may be a source of trust that influences drivers’ automation use decisions. This is generally examined in scenarios where the driver is attentive to the automation. However, future automation may not require driver attention. Thus, driving style similarity must be revisited as various levels of attention may affect how drivers accumulate evidence of driving style similarity. A model that considers two levels of driver attention (attentive and non-attentive) in the drivers’ evidence accumulation process is proposed. Dissimilarity is used as the evidence that contributes to driver distrust leading to automation disuse. A 2 (conservative and aggressive automation) × 2 (attentive and non-attentive driver) simulation of car following is examined where the driver is assumed to behave conservatively. Simulation results indicated no significant effect of driving styles on disuse, but inattentiveness mitigates disuse. These simulations may inform future empirical studies.

First impressions of a financial AI assistant: differences between high trust and low trust users

“…If one places too much trust in a financial advisor who actually does not deserve the level of trust, for example due to bad intentions or a lack of competence, this can lead to over-reliance, wrong decisions and associated harms for the trustor (Miller et al, 2016 ; Robinette et al, 2016 ; Chong et al, 2022 ). If, on the other hand, one distrusts a financial advisor despite their good intentions and high level of competence, one may miss out on possible benefits or have to allocate alternative resources to reach comparable results (Grounds and Ensing, 2000 ; Kamaraj and Lee, 2022 ).…”