The traditional approach to true artificial intelligence is to develop abstract mathematical functions, until a 'bug' happens, which would somehow enlighten the structure and endow it with self-consciousness - thus creating a new artificial life form.
Computer scientists are not the only ones having this problem; biologists tend to attribute the creation of biological life to a chanceful accident.
Psychology has a thing in common with computer science too; neither knows what intelligence is, so it is defined with the IQ and Turing tests, respectively. The problem with these kinds of definitions is that it is useful only in a limited context and does not offer anything to further the understanding of the idea of 'intelligence'.
Following a single line of thought, namely that combining abstract mathematical functions with what is seen as lifeless materia achieves at one point awareness/life, leads us to a form of machinistic determinism. To break out of this seeming impasse, we can come to two conclusions:
1. We have to break down reality, bundle it with the concept of 'consciousness', and build it back up. Thus, everything is 'intelligence' to a various degree.
2. We accept the determinism and the machine aspect of intelligence, and conclude there is no true intelligence (including human intelligence) and that the world is and illusion.
To the question "What is human?" the answer usually boils down to "Human is like me." I presume the question "What is intelligence?" yields a similar answer: "Intelligence is like me."
We could use wider perspectives and self-identity beyond personal egos.