10.10.25 - Fatemeh (Noosheen) Nazari

10.10.25 - Fatemeh (Noosheen) Nazari

🎙 Fatemeh (Noosheen) Nazari 👥 117 📅 October 13, 2025 ⏱ 41 min 👁 39 📄 original study 🧭 2026-08-16
Available in: English (current) Français

Keywords

autonomous vehiclessafetypublic perceptionsurveypolicy

Summary

The seminar presents a study on public perception of autonomous vehicle (AV) safety in dilemma situations, where AVs must choose between protecting passengers or pedestrians. The researcher, Noosheen Nazari, introduces the context of growing AV deployment and persistent public safety concerns. The study uses a survey conducted in San Francisco and San Antonio, collecting data on demographics, travel behavior, and attitudes. A dynamic Bayesian network model is employed to analyze how different scenarios (e.g., prioritizing pedestrians unconditionally, prioritizing based on law-breaking, children, or majority) affect confidence in AV safety and willingness to ride. Latent class analysis identifies four decision-making styles: detached, highly engaged, utilitarian, and balanced. Results show that unconditional pedestrian prioritization leads to sharp drops in confidence and willingness, while conditional rules cause moderate declines. San Francisco respondents show higher initial trust and greater sensitivity to scenarios. Policy insights suggest avoiding rigid rules and tailoring strategies to different user groups and regions. Future work includes generalizability and incorporating physiological data.

161 words

Critical Evaluation

Value of the Information & Strength of the Argument

The presentation provides valuable insights into public attitudes toward AV safety, using a rigorous survey-based methodology and advanced modeling. The argumentation is well-structured, clearly linking research gaps to the study design and findings. The use of latent class analysis adds depth by capturing heterogeneity in decision-making styles. However, the presentation is high-level, with limited discussion of model validation and statistical details, which may leave some questions about robustness.

Scientific Rigor, Source Quality, Title Accuracy

The speaker references prior work and mentions publications in Transportation Research Part C and D, but no specific sources are cited in the video. The title is simply the speaker’s name and date, which is typical for seminar recordings and does not reflect the content. The presentation appears scientifically rigorous, with a clear methodology and results, but the lack of direct references limits verification.

147 words

Title / Content Match

The title is simply the speaker's name and date, which does not convey the content, but this is typical for seminar recordings.

Quality & Reliability

7/10

The presentation describes a peer-reviewed study using a survey-based data-driven approach, with clear methodology and results. However, the video is a seminar recording with limited detail on statistical validation and no direct links to the underlying paper.

Key Moments

Cited Sources

  • AAA survey on autonomous vehicle concerns — Mentioned in the presentation as a source of statistics on public fear of AVs.
  • Transportation Research Part C and D publications — Speaker mentions her research has appeared in these journals, but no specific papers are cited.

Concurring Sources

  • Moral Machine experiment — Similar findings on public preferences for AV decision-making in ethical dilemmas.

Dissenting Sources

  • AAA survey on AV concerns — The survey indicates high public fear, but the study's findings suggest conditional acceptance, which may seem contradictory.

Contribution & Novelties

The study contributes a data-driven approach to understanding public perception of AV safety in dilemma situations, using a dynamic Bayesian network and latent class analysis to capture heterogeneity. It provides policy insights on how different decision rules affect trust and acceptance. The comparison between cities with different AV exposure adds a contextual dimension.

Pour aller plus loin :

  • Moral Machine experiment — A seminal study on public preferences for AV moral decisions.
  • Dynamic Bayesian networks — Overview of the modeling technique used.
  • Latent class analysis — Explanation of the method for identifying unobserved subgroups.

94 words

Radar Profile

The radar profile shows balanced scores across information quantity, quality, technical level, and reliability, indicating a well-rounded presentation with moderate depth and strong methodological grounding.

Reliability 7/10

💬 No comments were provided for analysis.