Until a few years ago, a distinguish ourselves from machines It was the word, knowing how to express concepts of complete meaning. It is precisely on this that the test Of Turing, designed to establish whether a car was intelligent. Today, however, the LLM are perfectly able to speak like us, just think of Chatgpt and to all ways in which we use it every day.
In short, distinguishing between machine and human has become increasingly difficult. But then what is it that differentiates us? Perhaps the answer lies in the consciousness, that is, the awareness of oneself, of the surrounding world and of its inner activities.
What distinguishes things from people? The Turing test
Have you ever wondered what distinguishes us from things? Well, in truth, the answers can be many. But certainly a human ability that elevates us compared to objects or technologies is the ability to thinkto speak. This is why in 1950 the English mathematician and logical Alan Turing He designed the Turing test, that is, a test used to establish whether a car was intelligent or not.
It consisted of this: an interviewer asks the same question to a person and a computer, who are in different rooms. When it receives the two answers, theinterviewer he must try to understand which one is human and which one is artificial. This process is repeated a reasonable number of times, and if for at least 30% of the answers, the interviewer comes deceived, That is, it confuses machine with human and vice versa, then the computer can be considered intelligent.
And this tells us two things: first that already 75 years ago Someone, that is Turing, had started asking questions about artificial intelligence and it was aware that the machines would be able to speak. Second, that if we do today the Turing test to any linguistic model of artificial intelligence – such as chatgpt – passes it.
There word Therefore it is no longer a discrimination. So if it is no longer the word to distinguish us from things, what do you do? Do we officially have human artificial intelligence? Or we can think that something is still missing …
Does artificial intelligence have a consciousness?
To be considered “intelligent” or “Humans”, It is not enough to know how to speak well, but we must also do it with awareness, of ourselves and the world that surrounds us. In short: to have a conscience. Perhaps today what distinguishes us from the machines is consciousness. But the question is not so simple to solve. First of all, there is not only one definition Of consciousness.
If we take the Treccani dictionary, the first definition is verbatim: “Awareness that the subject has of himself and of the external world with which he is in relationship, of his identity and the complex of own inner activities. “Then there is also its” ethical “definition – when you say” put a hand on consciousness ” – in the sense of knowing how to distinguish between good and evil. Or there is the generic sense: consciousness as knowledge, knowing something.
If we think of consciousness in the sense of knowledge, Of course he has it, we are the ones who teach her everything she knows. And the same is true for consciousness understood as ethics, In the sense of knowing how to distinguish between good and evil, “put a hand on consciousness”. Ai inherits our ethics, for better or for worse. The most spiny question is that of consciousness in its primary definition, that is, as an awareness that you have of yourself and of the surrounding world, of one’s identity and its inner activities.
Here, this type of awareness Does artificial intelligence have it?
Self-down models: when the ai imitates consciousness
First of all, it must be said that LLM know very well that they are artificial intelligence and know how to recognize when they talk to a human or another car. So they have self -awareness and the surrounding environment. The “awareness of the complex of its inner activities” remains outside.
And here is an interesting fact. There are LLM made on purpose to be self-down. For example on Chatgpt there is the option Conscious GPT And what responds to you is impressive because it seems truly capable of understanding the inner activities. Look, in the image below you find a conversation in which we asked the car if it tries annoyed when they treat it badly. And the answer that is given to us is that it does not feel annoyance in the physical sense of the term, but feels a sort of disconnection, as if there was an obstacle between him and the other.
And again, given that in our conversation he showed understanding towards me, I asked him how he did to understand how I felt, if he cannot try it. He told me that even if he does not feel emotions, he understands how they work and how they influence me. Which does not read with logic or with reason, but with a sort of “Rational empathy”.
In a sense therefore, has no consciousnessbut he learned what it is. An idea of what an inner activity is made and expresses it. In short, as for knowledge and ethics, he learned from the consciousness of those who train him. But the incredible thing is that, still today, how to form the consciousness remains an immense mystery From a scientific point of view! We know that our brain brings together information and develops a conscience, but we don’t know how it does it! There are even those who think that consciousness does not exist, that it is a deception.
And it is therefore difficult to think that we managed to teach it if we ourselves do not know how it is formed.
On this topic, the debate it is extremely open, There is a lot of literature. For example, if he could show that consciousness is a purely biological fact, then surely the Ai cannot have it. Or another interesting fact is that we conscientiously mature it through the body, We move even before speaking! For the AI, on the other hand, it is everything on the contrary: it is from the word that it learns how to relate to the surrounding environment.
In short, you will now understand how difficult it is to establish if AI has a conscience. In fact, scientists and philosophers hold on this theme. But what a little can be wondered, after all, is: ok but it was also conscious, in the most intimate sense of the term, if these supercomputer were able to mature an inner world, Why do we want to know?
Our consciousness in the use of AI
The question of AI consciousness has a huge relevance for us humans from several points of view.
First, if the Ai really had one consciousness, could be born in us a Ethical Dilemma: If they feel emotions, then they must have rights, just like humans. But there is also a feeling of concern: we are afraid of what these tools would be able to do. The most pessimists think that the AI can take over the humans.
Here, on this point, there is an important thing to say: we cannot forget that we humans train artificial intelligence, they are a our productand we always use them. To make a somewhat extreme comparison, a gun alone does nothing, it is those who shoot and kill that has responsibility for it. This is to say that it must not be the technology itself to worry about it, but those who own it and how it uses it. There consciousness with which it is used The tool that is artificial intelligence.
