The following incident is typical of today’s widespread blind – and nonsensical – belief in the intelligence of these systems. I had written an article in which I analyzed artificial intelligence and explained why these systems are not intelligent. Someone commented: “Of course such a system is an intelligence.” I politely asked for a definition of the term intelligence, because without one, a discussion about the intelligence of these systems is completely pointless. He ignored my request and tried to add weight to his opinion by referring to his professorship and his knowledge about the nervous system. I wrote again that we don’t get anywhere without a definition of intelligence. He evaded again, and I asked a third time what exactly he meant by intelligence. Then nothing more came. Anyone who claims that these systems are intelligent doesn’t know what intelligence is. There is no shame in that, because this shortcoming can be repaired. But most people don’t want to face this question – that’s how blind their belief is.
In the following, I explain what intelligence is and use two examples to show why the alleged AI systems are not only not intelligent, but really stupid.
The word intelligent comes from the Latin words inter (meaning ‘between’) and legere (meaning ‘to read’). So being intelligent means ‘reading between.’ Intelligence is the ability to deal with a situation when you don’t know what to do – because you have never learned it and have no experience. Intelligence is, therefore, shown above all in dealing with the unknown. That’s why creativity is an essential part of it. Knowing how to deal with a situation is not an expression of intelligence, but of education and functioning. Education and functioning are the opposite of intelligence.
Children are the most intelligent people on this planet because they do almost everything without having learned it. A computer system can only do what it has been taught, ie programmed. You can find out more about intelligence in my article “Why and How We Lose Our Innate Intelligence – and How We Can Regain It.”
Now to the first example.
One of the most demanding activities of a mathematician is proving. A proof is a chain of logical arguments that shows the validity of a statement beyond doubt – and does so with such precision and clarity that intelligent people can understand and confirm the steps.
In high school, proving is only taught using the simplest examples, if at all. In a mathematics degree course, proving is a central activity; and so it is in mathematical research.
For my doctoral thesis, I developed a computer system that proves geometric conjectures automatically. My system could prove theorems that in the past took mathematicians years to prove.
Most people will now say: this system must be intelligent. But it is not intelligent. As its author, I know that. The system does not do what a mathematician would do to prove a geometric statement. It translates the geometric proof problem into an algebraic computation task. This algebraic task is pure craftsmanship, but it is far too time-consuming to be done by hand. A computer usually only needs a few minutes, maybe only seconds. This computational work follows an algorithm that was created 60 years ago by the intelligence of a mathematician. No matter where you look with these systems: it’s always about the intelligence of the people who crated or enabled the system.
Here is the second example.
I came across an image that an AI system had produced. It showed the palm of a hand with hand lines. But the system had generated impossible lines, ie a line pattern that does not exist in any hand. Since psychological hand analysis is one of my areas of expertise, I immediately noticed this flaw. This example shows how stupid the system is.
If an even remotely intelligent person were to draw lines of a hand, they would first look at real hands. After just a few hands, they would see the pattern; they would see which lines there are in every hand and how they are arranged. Then they would either draw the lines of a specific hand, perhaps their own, or lines that correspond to the pattern. It would be foolish to draw lines that he had not seen in any hand. It would also be stupid to draw hand lines without looking at lines of real hands first. But the system has done just that. Even though it has access to countless hand images where you can see the lines, it produced impossible lines. That’s utterly stupid!
When I saw this example, I remembered my high school math teacher, who also taught us programming. He had said: “A computer is the stupidest thing you can imagine. If you tell it to leave the room, it will try to go through the closed door. You must tell it to open the door first. It's that stupid."
Obviously, knowledge of hand lines was not programmed. If I point out this omission to a programmer, he could improve the system; this would make it better – but not more intelligent. To improve, a computer system needs outside intervention – or it was programmed to learn in a certain way in certain situations. Every child grows on its own. It doesn’t need a teacher or a learning method it had been taught; it grows through self-reflection, which is an elementary aspect of intelligence. Self-reflection means checking the quality and meaningfulness of your actions, and learning from it.
Every tool, and therefore every AI system, only represents the intelligence of those who created it. It is not intelligent itself.
In a way, the alleged AI systems are an intelligence test for us. Anyone who refuses to talk about a definition of intelligence in connection with AI is demonstrating that they do not (or no longer) possess enough intelligence. The Greek philosopher Socrates said:
“The beginning of wisdom is the definition of terms.” (Socrates)
You can find a more in-depth analysis of artificial intelligence in my article “The Good, the Bad, and the Ugly of Artificial Intelligence.” There I also explain why most people confuse intelligence with education or the ability to learn and why the term ‘artificial intelligence’ is harmful.
Further reading:
Article “Why and How We Lose Our Innate Intelligence – and How We Can Regain It”
Article “The Good, the Bad, and the Ugly of Artificial Intelligence”
Well written. Intelligence is sadly misunderstood completely by most folks.
It is amazing how people agree to perceive AI as intelligent without even understanding the concept of intelligence , only because they are told so. Hiding behind AI are always humans; however, they prefer to yield responsibility because the programming is not always harmless and the creators pretend that the AI has its own intelligence that is out of their control