Planning and Policy for Safer Roads with Autonomous Vehicles: Moral Decision Making Behavior in Dilemma-inducing Situations

Nazari, Fatemeh; Noruzoliaee, Mohamadhossein · 2025 · ROSA P / Carnegie Mellon University. Traffic21 Institute. Safety21 University Transportation Center (UTC)

archive: archived pipeline: cataloged verified

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

Summary

This study investigates how public confidence in autonomous vehicle (AV) safety and willingness-to-ride respond to policy interventions regarding moral decision-making in crash-imminent scenarios. Motivated by the need to align AV regulatory frameworks with public expectations rather than purely philosophical prescriptions, the research examines how individuals navigate safety trade-offs, such as prioritizing pedestrians versus passengers. The authors aim to provide evidence-based guidance for policymakers and industry stakeholders by capturing the cognitive and attitudinal factors that shape public acceptance of AV behavior in complex, dilemma-inducing situations. The researchers developed a Dynamic Bayesian Network (DBN) framework to model the causal and temporal dependencies between baseline attitudes and post-policy perceptions. The DBN integrates socio-demographic factors, travel behavior profiles, and attitudinal indicators to simulate how policy scenarios influence confidence in AV safety and willingness-to-ride. Data were collected via an online survey administered in 2025 to residents of San Francisco (SF) and San Antonio (SA), yielding 176 and 159 valid responses, respectively. This dual-city design allowed for cross-regional comparisons between a technology-forward urban environment and a more car-oriented metropolitan area. The survey measured baseline attitudes, AV familiarity, and responses to four specific policy scenarios: unconditional pedestrian prioritization, passenger prioritization only when pedestrians break laws, prioritization of children, and prioritization based on group size. Results indicate significant city-level and attitudinal heterogeneity. SF respondents exhibited higher baseline confidence and willingness-to-ride but demonstrated greater sensitivity to policies prioritizing pedestrians, with significant declines in acceptance across all scenarios. In contrast, SA respondents showed more stable and modestly positive shifts, particularly when policies favored passengers. Stratified analyses revealed that pedestrian prioritization policies amplified existing differences across baseline confidence groups, whereas policies prioritizing children or group size promoted convergence toward mid-scale attitudes. Additionally, SF residents displayed higher AV familiarity and a more balanced mix of emotional and rational decision-making styles, while SA residents leaned toward instinctive decisions but maintained strong rational endorsements. The DBN model successfully captured these dynamics, showing that rule-based behaviors (e.g., penalizing law-breaking pedestrians) reassured respondents by signaling predictability. The findings underscore the value of DBNs in modeling the causal dynamics of AV acceptance and highlight that public responsiveness to moral algorithms varies significantly by geographic and socio-demographic context. The study concludes that AV decision-making frameworks must account for these heterogeneous public expectations to ensure social acceptance. By providing a scalable, empirically grounded tool for simulating policy impacts, the research offers actionable insights for designing AV systems that balance technical feasibility with widely held public preferences, thereby advancing the integration of autonomous vehicles into real-world transportation systems.

Key finding

San Francisco respondents exhibit greater sensitivity to policies prioritizing pedestrians, resulting in significant declines in AV acceptance, whereas San Antonio respondents show comparatively stable and modestly positive shifts, particularly when policies favor passengers.

Methodology

survey

Sample size: 335

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 bulk_ingest_rosap on 2026-05-23 (7 acquisition events logged).

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

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

• 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).

• Theoretical Contribution: computational model