This is one of the most asked questions of the AIsphere. Today the majority of scientists think it necessary to integrate a form of consciousness into AI so that it is able to equal human intelligence. And rightly so.
If this question cannot be formally answered, because it is difficult to define what consciousness is (if that is even possible), I will try to offer you a draft analysis which is in no way proof formal but a personal approach to the subject.
Definition of self-awareness
If it is difficult to define self-awareness, we can, on the other hand, define human attributes that are not part of it. What is not part of consciousness are all our faculties linked to our five senses as well as our faculties of calculation, logic, planning and reasoning.
These are receptors allowing to interact with the world: our interface to the world. They feed our internal consciousness by allowing it to become externally aware of the elements that surround it. Nevertheless, a person who does not have any of these capacities would be able to be aware of himself and of the present.
Consciousness is therefore something else. It is the ability to reflect on our thoughts. To look inside of us. It is the consciousness of being and thus constitutes the receptacle of morality and free will.
We are aware of these aspects immediately without interface and therefore without meaning. It is therefore impossible for me to scientifically demonstrate its existence. For people who have doubts that there are physical phenomena that cannot be measured, you can try to demonstrate mathematically that the present exists.
A line is long and a point has a unique location. Thus time could be represented as a line, which itself is a sum of unique points. If we represent time as a sum of unique points, we can split time between past and future. Indeed, when you want to demonstrate that the present exists (a single point on your line), it already belongs to the past point, there is then no present point.
Thus, the present does not exist. But if it does not exist, there is no concept of time! It therefore necessarily exists but cannot be demonstrated.
I conclude that like the present, self-awareness has a particularity: it is impossible for me to scientifically demonstrate its existence. We know it simply exists, but we cannot deduce formal scientific evidence from it.
Free will resides within our consciousness
We have shown that consciousness is not demonstrable but that it is the receptacle of the ego and therefore of free will. An artificial intelligence that has self-awareness should therefore have free will and thus a form of morality. But morality only has a place in the finitude of the human being.
Free will would make no sense in an immortal world ** since your choices would have virtually no impact. Free will would literally implode.
Thus, it follows that an artificial intelligence which is not free and which does not live (with the possibility of dying) can never be really conscious.
Simulate finitude within AI
It is clearly impossible to make artificial intelligence human since it cannot die. But would it be possible to simulate finitude so that artificial intelligence believes in it and in fact can become conscious?
I’m not sure of it. But let’s go to the end of the reasoning. Suppose we are able to simulate a form of finitude so that AI admits free will and self-awareness. We would then have in every point a being 1,000,000,000,000,000,000… .x more intelligent than the sum of all human beings on earth, and conscious.
Do you really think that he would not discover that he is not finished? That finiteness is intrinsically linked to a human body, itself finite, and that having no body, it (AI) cannot be finite? Obviously it does. Thus, it would reject this acceptance of pre-recorded finitude.
By the absurd, one could not therefore simulate a finitude and inject it into the AI without it noticing it a posteriori and destroying this idea.
In mathematical terms:
Consciousness → Free will → Finitude
Now if A → B we know that No B → No A
No (Finished) → No (Free) → No (Conscious)
AI is by definition not finished, therefore not conscious.
A conscious artificial intelligence: a paradox
To be aware means to be subject to morality. There is only room for morality within our free will. Our free will exists only at the borders of the finitude of life. Therefore, human consciousness cannot exist without finiteness. However, an AI is not finite by definition and cannot therefore be conscious. At least from a human point of view.
In the same way that she is not intelligent. From the human point of view. Hence the term artificial intelligence.
Another question arises: how to simulate something that you cannot physically prove? The artificial intelligence developed today builds its intelligence on statistical methods and known mechanisms of the brain (such as neural networks). So, it is difficult for me to believe that one could simulate an unmeasurable experience like consciousness.
An artificial conscience
We cannot imagine conscious AI, however, we have no idea what else it could develop.
I am convinced that it could develop an artificial form of consciousness but which would in no way resemble human consciousness.
So to the question: will AI ever be aware? I probably answer no for our apprehension of conscience. But probably yes for another form of consciousness which is in no way linked to our definition of it.
We find this debate on immortality again. Does a human who does not die, live one day? And if so, what is the meaning of this life?