Human and machine drivers: Sharing control, sharing responsibility.

Area of Science:

• Human-Computer Interaction
• Automotive Engineering
• Law and Technology

Background:

• Machines are gaining autonomy, complicating responsibility attribution when they cause harm.
• Automated vehicles (AVs) present unique challenges in assigning blame for accidents.
• The 2018 Uber AV crash highlighted issues with both human and machine driver performance.

Purpose of the Study:

• To investigate human judgments of responsibility for automated vehicle crashes.
• To examine the relationship between automation level and perceived human responsibility.
• To explore the mediating role of perceived human controllability in these judgments.

Main Methods:

• A cross-national online survey (N=1657) was conducted.
• Participants evaluated hypothetical automated vehicle crashes.
• The study manipulated automation levels, crash severity, and participant nationality (China and South Korea).

Main Results:

• A negative association was found between automation level and human responsibility.
• Perceived human controllability partially mediated this negative association.
• Findings held true across different responsibility metrics, nationalities, and crash severities.

Conclusions:

• Increasing automation in vehicles reduces human responsibility for crashes.
• Perceived human control is a key factor in assigning responsibility.
• Current driver-centric legal systems may need to evolve into control-centric frameworks for AVs.