The Tribunal · AI & Ethics · Clickbait Desk
A real-world case for AI ethics: Uber’s 2018 self-driving car fatality
A real-world case for AI ethics: Uber’s 2018 self-driving car fatality
In March 2018, in Tempe, Arizona, an Uber autonomous test vehicle — a Volvo XC90 — struck and killed 49-year-old pedestrian Elaine Herzberg. It became the first known pedestrian death caused by a self-driving car.
The vehicle’s system detected her about 5.6 seconds before impact, but it was not configured to brake for a jaywalker. The backup safety driver was also streaming a TV show at the time.
Uber was not criminally charged and later settled with the family. The backup driver pleaded guilty to negligent homicide in 2023.
This case raises a difficult ethics question: when an AI system fails in a high-stakes environment, how should responsibility be divided among the company, the human monitor, the designers, and the system itself?