Project

Back to overview

ARC - Autonomous Responsible Cars

English title ARC - Autonomous Responsible Cars
Applicant Shaw David
Number 190253
Funding scheme Spark
Research institution Institute for Biomedical Ethics University of Basel
Institution of higher education University of Basel - BS
Main discipline Sociology
Start/End 01.02.2020 - 31.03.2021
Approved amount 70'125.00
Show all

All Disciplines (2)

Discipline
Sociology
Information Technology

Keywords (6)

Self-driving cars; Smart cars; Decision-making; Accident ethics; Empirical ethics; Autonomous vehicles

Lay Summary (French)

Lead
Comment les voitures autonomes et autres véhicules intelligents devraient-ils prendre des décisions qui ont une incidence sur qui doit vivre ou mourir dans les accidents ? Le projet ARC proposera aux participants de prendre des décisions éthiques réfléchies et motivées sur une grande variété de cas d'accident à l'aide de types de données pertinentes.
Lay summary

Comment les voitures autonomes et autres véhicules intelligents devraient-ils prendre des décisions qui ont une incidence sur qui doit vivre ou mourir dans les accidents ? Certains suggèrent que les voitures devraient imiter les réactions instinctives des humains ou bien refléter les préférences de la société pour décider qui prioriser dans de tels incidents. Mais trois grandes lacunes existent dans ce domaine. Premièrement, les recherches actuelles ont négligé certains types de données qui pourraient être potentiellement utiles pour décider comment les voitures intelligentes devraient réagir en cas de collision. Deuxièmement, les recherches qui ont été menées se sont concentrées sur l'évaluation des réactions intuitives et morales du public, plutôt que sur des réponses réfléchies. Enfin, aucune étude sur l'éthique des véhicules autonomes n'a été menée en Suisse à ce jour, ainsi le comportement et les opinions morales subjectives des citoyens suisses n'ont pas encore été prise en compte au fond. Le projet ARC proposera aux participants de prendre des décisions éthiques réfléchies et motivées sur une grande variété de cas d'accident à l'aide de types de données pertinentes. Tout d'abord, une série d'entretiens avec le grand public sera menée au cours desquels leurs comportements en cas d'accident seront testées,  puis divers études de cas seront explorées. Deuxièmement, les résultats de la première phase serviront à concevoir une enquête expérimentale d'éthique à grande échelle auprès du public Suisse. Enfin, les résultats de l'enquête serviront à créer un référentiel éthique pour les voitures autonomes responsables.

Direct link to Lay Summary Last update: 04.12.2019

Lay Summary (English)

Lead
How should autonomous cars and other smart vehicles make decisions that affect who lives and dies in accidents? The ARC project will encourage participants to give reflective and reasoned ethical decisions on a wide variety of crash scenarios using expanded types of relevant data.
Lay summary

How should autonomous cars and other smart vehicles make decisions that affect who lives and dies in accidents? Some have suggested that car should mimic instinctive human responses, or reflect societal preferences on who should be prioritised in such incidents. But three significant gaps exist in this area of research. First, existing research has neglected potentially relevant data types that could also be used as inputs for decisions about how smart cars should react in crash scenarios. Second, the research that has been conducted has focused on assessing the intuitive moral reactions of members of the public, rather than considered reflective responses. Finally, no such research on the ethics of autonomous vehicles has been conducted in Switzerland, so the attitudes and subjective moral views of Swiss citizens have not yet been captured at a deep level. The ARC project will encourage participants to give reflective and reasoned ethical decisions on a wide variety of crash scenarios using expanded types of relevant data. First, a series of interviews will be conducted with members of the public in which their attitudes to crash scenarios will be tested and various case studies explored. Second, the results of the first phase will be used to design a large-scale experimental ethics survey of the public across Switzerland. Finally, the results of the survey will be used to create an ethical framework for autonomous responsible cars.

Direct link to Lay Summary Last update: 04.12.2019

Responsible applicant and co-applicants

Employees

Abstract

With the advent of automated and autonomous vehicles, much media attention and research has focused on the question of how these cars and other smart modes of transport should make decisions that affect who lives, dies and is injured in crash and other accident scenarios. Some have suggested that care should mimic intuitive human responses, or reflect societal preferences on who should be prioritised in such incidents. But three significant lacunae exist in this area of research. First, existing research has neglected potentially relevant data types that could also be used as inputs for decisions about how smart cars should react in crash scenarios. Second, the research that has been conducted has focused on assessing the intuitive moral reactions of members of the public, rather than considered reflective responses. The fact that people react instinctively while driving a care does not mean that their decisions are sound, or that such decisions should be used as a model for ethical decision making by computers that can process thousands more types of relevant input. Finally, no such research on the ethics of autonomous vehicles has been conducted in Switzerland, so the attitudes and subjective moral views of Swiss citizens have not yet been captured at either the shallow or any deeper level. The ARC project aims to generate the first data required to fill these three research gaps by using experimental methods to prompt research participants to give reflective and reasoned ethical decisions on a wide variety of crash scenarios using expanded types of relevant data. ARC will adopt a mixed methods approach. First, a series of 30-35 interviews will be conducted with members of the public in which their attitudes to crash scenarios will be tested and various case studies explored. A method of ‘experimental ethics’ will be used; when participants give their immediate reaction to a scenario, they will be asked to provide reasons for it, and further detail will be added to the scenario to establish whether their decisions are affected. The interviews will then be transcribed and coded into different themes and analysed. Second, the results of the first phase will be used to design a large-scale experimental ethics survey of the public across Switzerland, with a similar approach to that of the interviews but in survey form. Finally, the results of the survey will be used to create an ethical framework for autonomous responsible cars. ARC is a cutting-edge study in several ways. It is the first empirical study of attitudes to smart car crash decision-making in Switzerland; it is the first experimental ethics study of this type in the world; and it is the first study to consider the potential impact of different data types on ethical decision making in this context.
-