Artificial intelligence is having its moment of glory. The early setbacks were followed, at the turn of the 21st century, by spectacular advances which are not fully understood: artificial intelligence remains partly opaque. Worse: although it is progressing, the distance that separates it from its proclaimed objective – to reproduce human intelligence – is not decreasing. To dispel this enigma, according to Daniel Andler – philosopher and mathematician – we must confront a second one: that of human intelligence.

This is not reduced to the ability to resolve any kind of problem. It qualifies through a judgment the way in which we face the situations, whatever they may be, in which we find ourselves. Intelligence is an irreducibly normative notion, like ethical or aesthetic judgment, and this is why it is deemed elusive. An “intelligent” artificial system does not know the situations, but only the problems that human agents submit to it. It is only on this point that artificial intelligence can support us. In fact, it solves an ever-increasing variety of pressing problems. This should remain its objective, rather than the incoherent one of seeking to match, or even surpass, human intelligence. Humanity needs docile, powerful and versatile tools, not pseudo-people with an inhuman form of cognition.

Reference 

Artificial intelligence, human intelligence: the double enigma

Daniel Andler

ISBN-13 ‏ : ‎ 978-1773271439