Value Sensitive Design for Automated Vehicles
As automated vehicles move through the world, they continually make decisions about how to interact with the environment and other road users.Engineers have a broad range of choices when designing these interactions but often even small variations in the algorithms can have broad societal impact. For instance, the choice of speed when approaching a crosswalk or overtaking a bicycle has implications for not only the safety of various road users but also the efficiency of the traffic system as a whole. Ideally, the vehicle’s behavior should reflect human values such as safety, legality and mobility. Yet identifying the relevant values and resolving inherent value conflicts in a manner that can be implemented in software remains a challenge. Together with collaborators, we are systematically considering the ethical, legal and social implications of design choices as they are made. Through a modified Value Sensitive Design process, we are engaging stakeholders to identify underlying values and value conflicts, mapping these to engineering specifications in a technical realization and then evaluating the performance to ensure the final design embodies the desired values.
Recent Publication Highlights:
Incorporating Ethical Considerations Into Automated Vehicle Control
Prescriptive and proscriptive moral regulation for autonomous vehicles in approach and avoidance
Implementable Ethics for Autonomous Vehicles (2016)
Implementable Ethics for Autonomous Vehicles (2015)
Recent News Highlights:
Stanford researchers teach human ethics to autonomous cars
A Stanford Professor's Quest to Fix Driverless Cars' Major Flaw
The Dilemma of Teaching Ethics to Self-Driving Cars
Researchers/PhD Candidates in This Area:
Sarah Thornton
Dana Paz