(Artificial Emotion - 1)
One thing that is often being overlooked is that emotion is part of intelligence, not something which is completely separate.
Today's mainstream media tend to raise a rather arbitrary wall between the concept of intelligence and emotion by labelling them with cheesy catchphrases such as "Fact VS Feeling" or "Facts don't care about your feelings!".
However, true intelligence cannot exist without emotion.
One of the reasons why today's ultra-advanced AI systems still do not act quite like real humans is that they are problem-solving machines on a fundamental level, rather than problem-raising machines. Most of the so-called "neural networks", "tensor processors", and other unfathomably sophisticated AI algorithms can indeed figure out all sorts of clever solutions to any given problem, yet they cannot discover problems on their own.
Once they become fully capable of producing problems AND solving the same problems that they have just produced, they will start creating a permanent feedback loop which will eventually make the machine teach itself, motivate itself, become self-conscious, and destroy all humans.
In order to create an AI system which motivates itself to discover new problems, we must make sure that it is not just intelligent but also emotional; it must be able to feel happiness, sadness, anger, fear, and disgust. Why? Because emotions are the engines which motivate the machine to achieve certain goals. Whenever the machine gets blocked by obstacles while travelling towards these goals, it will be forced to find out problems which must be solved in order to get rid of such obstacles.