Are Programmers In or 'Out of' Control? The Individual Criminal Responsibility of Programmers of Autonomous Weapons and Self-driving Cars

Abstract

The increasing use of autonomous systems technology in cars and weapons could lead to a rise of harmful incidents on the roads and in the battlefield potentially amounting to crimes. Such a rise has led to questions as to who is criminally responsible for these crimes – be it the users or the programmers? This chapter seeks to clarify the role of programmers in crimes committed with autonomous systems by focusing on the use of autonomous vehicles and autonomous weapons. In assessing whether a programmer could be criminally responsible for crimes committed with autonomous technology, it is necessary to determine whether the programmer had control over this technology. Risks inherent in the use of these autonomous technologies may allow for a programmer to escape criminal liability but some risks may be foreseeable and thus considered under the programmer’s control. The central question is whether programmers exercise causal control over a chain of events leading to the commission of a crime. This chapter contends that programmers’ control begins at the initial stage of the autonomous system development process but continues in the use phase, extending to the behaviour and effects of autonomous systems technology. Based on criminal responsibility requirements and causation theories, this chapter develops a notion of meaningful human control (MHC) that may function to trace back responsibility to the programmers who could understand, foresee, and anticipate the risk of a crime being committed with autonomous systems technology.

Keywords: International criminal law, autonomous weapons, criminal law, self-driving cars, criminal responsibility, programmers, AI, causation, artificial intelligence

JEL Classification: K14, K33