Licensing Autonomous Cars

This post is about licensing partly autonomous cars. Not legally speaking, autonomous cars are those which can drive themselves without any human input, and partly (or mostly) autonomous cars have reduced human input. They use sensors to provide electronic input to a rapidly interpreting computer, and the computer controls some or all of the car’s control mechanisms. There is a spectrum of automation: some commercially available cars offer semiautonomous parallel parking, others can keep a deemed safe distance behind the car in front, and the most advanced research and development (R&D) cars can drive across vast systems of roads without any notable human intervention. However, driving in “critical” situations, such as in extremely busy downtown streets with very high numbers of dynamic objects (bikers, pedestrians, other vehicles), has overwhelmed even the leading R&D autonomous cars’ computational systems, requiring human intervention. Power steering and anti-lock braking systems are early forms of automation.

 
Before diving into the the discussion, I will mention that most of the information in this article is from personal conversation with a scholar on the subject, Bryant Walker Smith, a fellow at the Center for Internet and Society at Stanford Law School and the Center for Automotive Research at Stanford, who also teaches a course on the subject: Legal Aspects of Autonomous Driving. He has been very kind in elucidating this undeveloped yet exciting field. He also has a blog on the subject.