Why Artificial Intelligence Is Simply Impossible
Trevor*

-Op-Ed-

BRUSSELS — At a primary school, young children were asked to complete the following sentence: “The cat has … legs and the bird … .” Conscientiously, the pupils filled in the blanks with the numbers four and two. That is, all but one of them, who instead wrote, “The cat has pain in his legs and the bird is sad about it.”

Does that one student’s answer indicate that he is somehow less intelligent than the rest? Of course not. For sure, he has a different intelligence, one that’s more unusual. He doesn’t think in the same way as the rest of his classmates, but that doesn’t mean he doesn’t think just as well, or even better.

So-called “intelligence” tests offer little information about actual intellectual capacity. It’s a bit like trying to estimate the value of a house by looking at what’s in the fridge. These tests date from a time when intelligence was for the most part limited to the logical and mathematical ability to calculate, rank, extrapolate or deduce. IQ is far from the perfect measure of intelligence because there are many kinds of intelligence. Intelligence is like blood type: There isn’t any that’s better than the rest, but some are more commonly shared. The question of one individual’s level of intelligence therefore matters less than that the type of intelligence.

Thankfully, since the 1980s and the emergence of work from academics such as American psychologist Howard Gardner, the plurality of intelligences and the necessity to combine them are no longer contested. Even better is that there are now many theories about the many forms of intelligence! Gardner himself has had to add one to his original list.

In addition to deduction, and mathematical and logical capabilities, there are, among others: musical intelligence (sensitive to sounds and rhythms), bodily intelligence (which frees up the full potential of all body parts), relational or emotional intelligence (the ability to identify the feelings and intentions of others), visual intelligence (which enables us to visualize before we build something and move objects into space), and language intelligence (the ability to think with words).

[rebelmouse-image 27089675 alt=”” original_size=”1024×683″ expand=1]

Unbeatable at table tennis, though — Photo: Global panorama

Almost none of these functions can be programmed so that a machine could carry them out. A computer can recognize a face, but it can’t find it beautiful. A computer has memory, but it can’t have memories. It can produce images, but it doesn’t have any imagination. A computer can learn from its mistakes, but it cannot regret them. It can compare ideas, but it cannot have an idea.

What we call “intelligence” isn’t a unique ability, but instead a set of skills, innate or acquired, that requires from us both to know and to ignore, to become emotional and to be detached, to ask questions and to answer. These skills are inseparable from surprises, sensations, intuition, laughter.

The very essence of intelligence is that it’s human and that it can’t be recreated by something artificial. If it became artificial, it would mean that we’d have given up on using our own. But the topic is raised again and again in the media. As soon as a computer defeats a human being at one game or another, the myth that artificial intelligence will become part of our lives is resurrected.

Computers can free us from many tedious chores, but that doesn’t mean they’ll make us free. They can help us foresee but not want. They can help us find information, but they won’t tell us what to look for. They can analyze the way things are headed, but they can’t understand what it means.

*Luc de Brabandere is a Belgian-born mathematician and senior advisor at the Paris office of Boston Consulting Group.

Translated and Adapted by: