“Artificial intelligence, human intelligence: the double enigma”, by Daniel Andler, Gallimard, “NRF essays”, 434 p., €25, digital €18.
The appearance around 2018 of major language models authorizing a computer to produce text, such as the very recent launch of ChatGPT or Bard, the most spectacular consequence of this discovery, have certainly changed the landscape of artificial intelligence (AI) . But has the ability to dialogue thus conferred on an automaton really given flesh to the Faustian dream that has accompanied the birth of AI since the end of the Second World War: to equal, surpass, or even replace man?
As evoked by the criminal metamorphosis of the HAL computer in 2001: a space odyssey, by Stanley Kubrick (1968), the prospect nourishes as much anxiety as enthusiasm. The philosopher Daniel Andler, one of the best specialists in cognitive science, opportunely publishes, in this hour of glory of technosciences, Artificial intelligence, human intelligence: the double enigmadense sum, effective, because full of knowledge and humor, which will know how to sober the anxious as the unconditional ones.
While he does not dispute the importance of the qualitative leap thus made, marked by the emergence of a “ethical awakening” in a discipline which until then had only been interested in intellectual prowess, Daniel Andler reverses current opinion and a whole philosophical trend fascinated by the growing blurring between the cyborg and the human, by demonstrating that the gap between technology and man, far from being fulfilled, has increased. Through a valuable reconstruction of the history of AI, which occupies the first part of the essay, he shows how the reality of AI’s progress and failures contradicts the “rhetoric of inevitable victory”. This triumphalism specific to the culture of “numerisphere” (a population of programmers and entrepreneurs that thrives alongside researchers) is not unlike the prophecies about the victory of the proletariat or the speeches confusing liberal capitalism with a law of nature, with no possible alternative or thinkable.
However, the “promethean dream” of an artificial intelligence that would join that of man, not only did not succeed, but above all, establishes Andler, is akin to a “chimera” or an ill-posed problem. This failure has conceptual and not technical causes. It is based on a restrictive definition of human intelligence as the ability to ” problems solving “ (where the machine may have the advantage). In reality, human intelligence, more and more conceived as mixed with affects, fulfills many other tasks (understanding the world, knowing the environment, etc.).
You have 40.51% of this article left to read. The following is for subscribers only.