Document de travail

How safe is safe enough? Psychological mechanisms underlying extreme safety demands for self-driving cars

Azim Shariff, Jean-François Bonnefon et Iyad Rahwan

Résumé

Autonomous Vehicles (AVs) promise of a multi-trillion-dollar industry that revolutionizes transportation safety and convenience depends as much on overcoming the psychological barriers to their widespread use as the technological and legal challenges. The first AV-related traffic fatalities have pushed manufacturers and regulators towards decisions about how mature AV technology should be before the cars are rolled out in large numbers. We discuss the psychological factors underlying the question of how safe AVs need to be to compel consumers away from relying on the abilities of human drivers. For consumers, how safe is safe enough? Three preregistered studies (N = 4,566) reveal that the established psychological biases of algorithm aversion and the better-than-average effect leave consumers averse to adopting AVs unless the cars meet extremely potentially unrealistically high safety standards. Moreover, these biases prove stubbornly hard to overcome, and risk substantially delaying the adoption of life-saving autonomous driving technology. We end by proposing that, from a psychological perspective, the emphasis AV advocates have put on safety may be misplaced.

Mots-clés

autonomous vehicles; automation; algorithm aversion; safety; illusory superiority;

Remplacé par

Azim Shariff, Jean-François Bonnefon et Iyad Rahwan, « How safe is safe enough? Psychological mechanisms underlying extreme safety demands for self-driving cars », Transportation Research Part C: Emerging Technologies, vol. 126, n° 103069, mai 2021.

Référence

Azim Shariff, Jean-François Bonnefon et Iyad Rahwan, « How safe is safe enough? Psychological mechanisms underlying extreme safety demands for self-driving cars », TSE Working Paper, n° 21-1215, mai 2021.

Voir aussi

Publié dans

TSE Working Paper, n° 21-1215, mai 2021