Explanations and Expectations

Haspiel, Jacob; Du, Na; Meyerson, Jill; Robert, Lionel; Tilbury, Dawn M.; Yang, X. Jessie · 2018 · OpenAlex

DOI: 10.1145/3173386.3177057

archive: archived pipeline: cataloged

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper addresses the challenge of building driver trust in automated vehicles (AVs), a critical factor for the acceptance of autonomous driving technologies. While drivers often distrust AVs due to unpredictable actions, research suggests that providing explanations for AV behavior can mitigate this distrust. The study specifically investigates the importance of the timing of these explanations, hypothesizing that providing an explanation before an AV takes an action will foster greater trust than providing it afterward. Additionally, the authors test whether giving drivers the option to approve or disapprove of an AV’s proposed action (in addition to the explanation) further increases trust. The research employed a within-subjects experimental design with eight participants (average age 25.7 years) who experienced four distinct driving conditions in a high-fidelity vehicle simulator at the University of Michigan Transportation Research Institute. The simulator featured a full Nissan Versa sedan and an SAE Level 3 automated driving system. The four conditions were: (1) no explanation; (2) explanation provided 7 seconds before the AV action; (3) explanation provided within 1 second after the action; and (4) explanation provided 7 seconds before the action, coupled with a driver option to approve or disapprove the action. Each condition involved a 6- to 8-minute drive containing three unexpected events (e.g., police vehicles, re-routes) that forced the AV to take unanticipated actions. Data collection included subjective measures of trust, cognitive workload, anxiety, and driver preference, as well as physiological data such as heart rate and galvanic skin response. Preliminary results from the eight participants indicated that the condition where the explanation was provided before the action (DC2) yielded the highest trust scores, both in direct attitude ratings and in rank-order comparisons. This condition also resulted in the highest driver preference and the lowest cognitive load. In contrast, the condition combining pre-action explanation with driver control (DC4) produced the lowest anxiety but the highest cognitive load. Notably, the addition of driver control did not appear to increase trust beyond the level achieved by the pre-action explanation alone. The authors explicitly state that no statistically significant results were found due to the small sample size, and therefore they refrain from drawing definitive conclusions. The significance of this work lies in its contribution to the design of human-machine interfaces for AVs. The observed patterns suggest that the timing of transparency is crucial; proactive explanations may be more effective for trust building than reactive ones. Furthermore, the findings imply that granting drivers control over AV actions may increase cognitive workload without necessarily enhancing trust, suggesting that interface designers should prioritize clear, timely explanations over complex control mechanisms. These insights aim to inform the development of AV systems that better align with driver expectations and promote acceptance of automation.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via openalex_abstract on 2026-05-08 (11 acquisition events logged).

StageOutcomeToolModelPromptAttemptsCompleted
discover success 1 2026-05-05
archive success core_acuk 53 2026-08-22
extract success cached 6 2026-08-22
clean success clean 2 2026-06-04
chunk success chunk 2 2026-06-04
embed success embed Qwen/Qwen3-Embedding-8B 1 2026-06-04
enrich success openalex 3 2026-05-08
promote success 2 2026-06-04
summarize skipped cached summ-v5 7 2026-08-23
tag success vector_similarity 15 2026-06-11

Summary generated by qwen3.8-27b-gittensor on 2026-08-22; verification: pending re-verification.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.