Machine Souls — Part 1: Introduction

Recently, we have made large strides towards AI. Programs based on deep learning have defeated world-class Go players, are assisting judges in the court of law, synthesize speech, and many more astounding achievements. As AI is becoming increasingly capable of performing tasks that have until recently been believed to achievable only by a human, the separation between man and machine is becoming less and less sure. The question of whether a machine can ever have consciousness or a soul is demanding the attention of more than just science fiction writers. That question can be reduced to the age-old question of what a soul and consciousness are: if we know what it means to have a soul and a consciousness, we can say if an entity has those. The practical relevance of this ancient question in today’s world is what equips us with new tools to answer it. In particular, because we can now build machines that are very similar to humans in some aspects, we can modify them to experiment with the various aspects of what the soul and consciousness are believed to be.

I base my approach on the assumption that anything that happens inside our universe follows physical laws, discovered and undiscovered. For example, if nothing in the universe can move faster than light, then souls with unfinished business cannot haunt a person in New York at 2 pm, then haunt a
person in Los Angeles at 2 pm and 11 milliseconds (it takes about 13 milliseconds for light to travel from New York to Los Angeles, assuming for this example that souls cannot travel straight through the earth).

My ultimate goal is to determine whether machines can ever reach the point of being considered to have a soul and the same level of sentience as humans. I first present a modern operational definition of soul that sheds some of the mystical baggage attached to the notion of a soul: a soul is merely the
influence that a person or an object has on the world; the amount and complexity of information that the person or object emits into the world. I then emphasize the distinction between having a soul and being sentient. Namely, an entity can be sentient only if it has complex computing unit (like a brain or a processor) that is programmed in such a way as to allow the entity to be aware of its existence, whereas a soul (information) can wander through the physical world and not be aware of itself. I then argue that there is no reason a machine cannot have a soul and be sentient.

read original article at——artificial_intelligence-5