The study, published in the journal Nature Machine Intelligence, conducted surveys with over 2,500 participants from the United States and Germany. The researchers presented people with hypothetical scenarios where an autonomous or a human-driven car became involved in an accident, and asked them to assign responsibility for the crash.
The findings highlighted a significant difference in public opinion on liability depending on the circumstances of the accident. When an autonomous car made an error, people tended to prioritize protecting the occupants of the vehicle and blame the car's manufacturer. However, when a human driver was at fault, participants often focused on penalizing them and gave less weight to potential harm caused to others.
"We saw a clear trend where people applied a 'pro-occupant bias' to autonomous cars. They favored shielding those inside the car, even if it meant increased potential harm to pedestrians and other road users," explains Dr. Johannes Himmelreich, lead author of the study and researcher at the Max Planck Institute for Human Development.
This bias aligns with the widely held belief that manufacturers and developers of autonomous vehicles should bear primary responsibility for their safety. However, legal experts point out that such an approach could discourage development and innovation in the autonomous car industry.
Furthermore, the study revealed a lack of consistency in moral judgments, even within the same individual. For instance, participants applied the pro-occupant bias more strongly in cases where the autonomous vehicle was transporting children.
"Our findings call for a more refined approach to safety and liability that considers the specific context of each accident and weighs a broader array of ethical considerations," says co-author Dr. Christopher Graeber from the University of Oxford's Future of Humanity Institute. "Simply putting strict liability on car manufacturers or drivers isn't enough."
The inconsistent public preferences pose a significant challenge for regulators and policymakers who are developing legal and ethical guidelines for autonomous cars. If liability is assigned based solely on the autonomous vehicle's behavior, it could fail to account for cases where human error or external factors contribute to accidents. Conversely, focusing solely on traditional liability rules for human drivers may overlook the unique attributes of self-driving vehicles.
"Our study shows just how complex the ethical and legal considerations around autonomous vehicles are," says co-author Prof. Dr. Christoph Trautman from the University of Vienna. "There's no straightforward solution, and public opinion is divided. Ultimately, it will be up to policymakers to strike a balance between innovation, safety, and fair liability."
As driverless cars become a more prominent feature on our roads, addressing the ethical and legal challenges they present will require careful consideration of all stakeholders, including auto manufacturers, legislators, and the public.
