“Delegation, Relinquishment and Responsibility: The Prospect of Expert Robots” draft in progress.
The article was written for the first “We, Robot” conference held in Miami in 2012 in collaboration with my favorite philosopher-engineer-guitarist and all around renaissance dude, Jason Millar.
Together, we question what role humans will occupy once robot experts are capable of performing a multitude of tasks traditionally delegated to human experts. We begin by describing IBM’s Jeopardy! winning robot Watson and Isaac Asimov’s ‘The Evitable Conflict,’ and argue that we are on the precipice of having to decide whether to relinquish some control to expert robots. In Part II of this paper we describe Watson’s development into a go-to-medical expert to demonstrate that there are already many instances where knowledge and control are being relinquished to machines. In Part III, we specify the kinds of robotic systems that we are concerned with, mainly those in which unpredictability in operations is a feature and not a bug. In Part IV we question when a robot might be considered an expert and in part V we examine the effect of expert robots on the question of relinquishing control of expert decision-making to machine systems. In Part VI we set out two situations in which humans will be working alongside robots and consider what might happen when disagreements between the two might arise. Finally we question how we might assign liability when an expert robot malfunctions.