Uncanny Moral Behavior

In trying to imitate human moral behavior, will we venture into the moral uncanny valley where the decisions appear superficially moral and yet disgust us? The uncanny valley describes a feeling of eeriness and ominousness which accompanies near-human like appearances and behavior in robots. As the sophistication of physical and virtual actroids has increased, under controlled circumstances they can be mistaken for humans. But outside of movement and appearance, I hypothesize here that as robotic, transhuman, or artificial entities behave more explicitly in a functionally moral manner humans will respond with feelings of revulsion. The paper describes several examples of functionally moral behavior and for each plot acceptability versus culpability. I assert that the most human-similar moral behavior is that which is identical in terms of culpability. This paper’s hypothesis is that when systems are very similar to humans in terms of culpability, their actions may become repugnant. In addition to traditional cases like industrial robot, corpse, zombie, and healthy person we will consider autopilots, fire control systems, an organ-swapping robotic hospital, an automatic trolley switch, and automatic quarantine nanobots. In these cases, some drawn from actual incidents, others from speculation we find it useful to draw parallels between the culpability of a human versus progressively more sophisticated moral systems. The argument is wrapped up through examination of the questions this model poses, such as whether a transhuman or cyborg who fuses human and artificial moral deliberation is less or more repugnant than an ordinary human. Or whether moral deliberation can be meaningfully distilled down into emotive notions like acceptability versus repugnance.