Reliable and transparent in-vehicle agents lead to higher behavioral trust in conditionally automated driving systems
DOI: 10.3389/fpsyg.2023.1121622
archive: archived pipeline: cataloged verified
Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)
Summary
This study investigates how system reliability and information transparency influence driver trust and performance in conditionally automated vehicles (CAVs). The research addresses the challenge of calibrating driver trust, which is critical for safe human-automation collaboration but difficult to manage due to dynamic system reliability and uncertain optimal transparency levels. Specifically, the authors examine how the presentation of transparent information via in-vehicle agents (IVAs) affects behavioral trust—measured as compliance with takeover requests (TORs)—and driving performance under varying reliability conditions. The researchers conducted a driving simulator study with 30 participants using a 2 (Reliability: low vs. high) × 2 (Transparency: proactive vs. on-demand) mixed factorial design. Reliability was manipulated by varying the accuracy of information provided by the IVA: the high-reliability condition featured 100% accurate information, while the low-reliability condition featured 67% accuracy. Transparency was manipulated through two agent types: a "proactive" agent that provided both "what" (intention/environment) and "why" (analytics/task) information simultaneously, and an "on-demand" agent that provided only "what" information initially, allowing drivers to request "why" information if desired. Participants completed two driving scenarios involving highway and urban roads, encountering takeover events triggered by system limits. Objective measures included TOR compliance rates, takeover time, and post-takeover vehicle control metrics. Subjective measures included trust, workload, and agent preference ratings. Results indicated that the on-demand agent was perceived as more habitable and faster in response than the proactive agent. Crucially, system reliability significantly impacted behavioral trust and performance. Drivers in the high-reliability condition complied with TORs more frequently than those in the low-reliability condition, but this effect was significant only when using the on-demand agent; no significant difference in compliance was found for the proactive agent across reliability levels. Furthermore, drivers experienced shorter takeover times in the high-reliability condition compared to the low-reliability condition, regardless of the transparency type. The reliability manipulation was successfully validated, with participants accurately perceiving the accuracy differences between conditions. The findings suggest that reliable and transparent in-vehicle agents lead to higher behavioral trust in CAVs. The study highlights that the method of information delivery interacts with system reliability to influence driver behavior. Specifically, on-demand transparency allows drivers to better calibrate their trust based on system reliability, whereas proactive transparency may mask reliability issues or cause information overload, leading to similar compliance rates regardless of accuracy. These results provide practical design guidelines for IVAs, suggesting that adaptive, on-demand information presentation can help mitigate the negative effects of low system reliability and improve driver performance and trust calibration in future automated vehicles.
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.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | Crossref | — | — | 1 | 2026-06-11 |
| archive | success | canonical_url | — | — | 1 | 2026-06-25 |
| extract | success | cached | — | — | 2 | 2026-06-25 |
| clean | success | clean | — | — | 1 | 2026-06-11 |
| chunk | success | chunk | — | — | 1 | 2026-06-11 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-11 |
| promote | success | — | — | — | 1 | 2026-06-11 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 1 | 2026-06-25 |
| tag | success | vector_similarity | — | — | 6 | 2026-06-11 |
| verify | success | — | — | — | 1 | 2026-06-26 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-25; verification: verified.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
- trust calibration
- automation
- automation surprise
- trust in automation foundations
- takeover transitions
- acceptance adoption
Information type
What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).
- Empirical Findings: self report data
- Theoretical Contribution: computational model, conceptual framework