Dalit Ken Dror Feldman

The Center for Cyber Law & Policy

Dalit Ken-Dror Feldman holds LL.B. (Summa Cum Laude) and B.Sc., Computer Science (Magna Cum Laude) Haifa University (2001); LL.M .in Commercial Law (Magna Cum Laude), Tel-Aviv University (2008); Ph.D. in Law, Haifa University (2018). Her research interests focus on Law & Technology and Legal Ethics, including Regulation of Technologies such as Artificial Intelligence and Unmanned Aerial Vehicle, Cyber Security and Cyber Terrorism, Intellectual Property and Open Access, Software Law and Information Law.

Dalit Ken-Dror Feldman is the legal supervisor of the Law, Technology and Cyber Clinic at the Haifa Center for Law and Technology, Faculty of Law, University of Haifa and a Lecturer at Zefat Academic College.

Abstract:
Autonomous Vehicles – Criminal Liability Challenges
Autonomous vehicles differ in degrees of automation. SAE (Society of Automotive Engineers) international, suggested in 2014, 6 levels of driving automation. These levels are affected by dependence on human driving. For example level 0 refers to no automation at all whereas level 1 is referred to "driver assistance". In levels 0-2 a human driver monitors the driving environment whereas in levels 3-5 automated driving system monitors the driving environment. When we talk about Autonomous vehicles we usually refer to level 3 and above.

Autonomous vehicles are based on artificial intelligence and machine learning. Therefore we assume that the more hours a car accumulates on real roads, the better it performs. During its training process, as well as in real life, autonomous vehicles might be involved in car accidents, as indeed has happened already.

On January 20 2016, a driver of a Tesla car was killed in China when the car crashed the back of a road-sweeping truck. On May 7, 2016 a driver of a Tesla car was killed in an accident when the car was in self-driving mode. Again on March 2018 we witnessed another deadly Tesla car accident in California. However, during 2018 we witnessed a new kind of autonomous vehicle accident. This time the victim was a pedestrian who crossed the road in Arizona. An Uber car was involved in that accident.

These car accidents clarify well why we should find a legal solution to criminal liability where a criminal offense is committed by an autonomous vehicle. In order to impose criminal liability, the existence of two elements is required: (1) the existence of criminal conduct; (2) general knowledge or intent with respect to the behavioral element (except in strict liability, for example).

In addition, with regard to the first element, sometimes proof of a certain result is also required. In the research, I examined several subjected models of criminal liability concerning artificial intelligence and four possible categories of criminal offenses in which an automated vehicle might be involved: (1) offenses of strict liability; (2) driving under the influence of alcohol; (3) negligent driving; (4) vehicular homicide. As a test case I referred to the Israeli Criminal Law. The research was conducted with the help of the Law, Technology and Cyber Clinic, University of Haifa, Faculty of Law, and led by Dr. Dalit Ken Dror Feldman.