Why ChatGPT is a charade

ChatGPT, the latest in artificial intelligence, is causing a lot of ink to flow. Much of the ink of those newspapers whose death he announces, being able both to provide solutions to all the problems, but also to comment on them, carrying out the work traditionally devolved to the press. Cathedral of algorithms built in the cement of the total information of the Internet, ChatGPT understands the questions it is asked and structures its answers according to conversational standards, with articulated paragraphs, each expressing an idea, addressing in turn the different aspects of a question to arrive, using humor on occasion, on nuanced conclusions.

Does ChatGPT mark a new step towards thinking robots?

Fascinating experience of seeing a machine, that is to say a piece of matter, explain to us what to think. “When the universe crushed him, Pascal said, man would be even more noble than what kills him, because he knows he is dying, and the advantage that the universe has over him, the universe knows nothing about. » ChatGPT crushes the man explaining why.

However, this would confuse information and knowledge. Knowledge is not only the fact of possessing information about the world, for example the name of the capital of Brazil, the mechanism of digestion, the reason for the war in Ukraine or the speed of light. It is get the idea why we ask these questions, that is to say, to consider the existence of a true ultimate, of a complete puzzle, of which the particular questions would be the pieces, allowing us to grasp the world in its principles.

Knowledge, in other words, is not a “bureaucracy of appearances” (Etienne Klein): it does not only consist in making the information comprehensible, but in making hypotheses on the puzzle itself, the nature of truth, and therefore in imagining rather than synthesizing, creating rather than popularize.

computer devices

This confusion has a cause and a driving force. The cause is the appearance of machines that seem produce ideas when they only operate calculations according to our parameters – in the manner of spontaneous generation, which claimed, before the Pasteurian discovery of microbes, that life was born in containers which already contained it imperceptibly. Thus, the “intelligence” of ChatGPT’s responses is already in its miniaturized algorithms. Will a blackboard be said to be intelligent because a mathematical equation has been written on it with chalk?

It will be argued that artificial intelligence produced the equation on its own, by combining mathematical data, from which we ourselves were unable to deduce anything. It is precisely this extraction of hidden truths that establishes the claim to the intelligence of our computer artifices. Except that it is only a question of combining data, and not of producing ideas, that is to say reflections on what is a data.

“Talking with ChatGPT is like chatting with someone very cultured, but completely crazy”

The attribution of a reflection to this calculating form also has a driving force: anthropomorphism, which lends human characteristics to what we do not understand. These machines seem to us all the more endowed with the alchemical power to transmute information into knowledge, as, ignoring their operating principle, we project our thoughts onto them, like in another era the divine intentions on unexplained natural phenomena.

Masquerade encouraged by the personification from ChatGPT’s answer. Because its novelty is less to synthesize content, which we already knew how to do very well, than to answer the questions put to it “in natural language”. The eyes of God looked at us through the mystery of nature, the intelligence of artifice through the simulation of human speech.

Artificial intelligence is misnamed

Borges had imagined a library that would contain all possible books, by random and exhaustive combinations of all possible words and phrases. A writer would thus only “choose” his book from the immensity of writable books. Similarly, a game of chess is only the choice, at each branching of moves, of one game among the multitude of possible games. Let’s go further: isn’t our whole life the choice of a real life among all possible lives? And in the already written totality of the possible, doesn’t ChatGPT show us the best of them?

ChatGPT 3 has no answer to social inequality

Confusion again. We would take the thing for the reason for the thing, the sign for the idea, the past instants for the creative duration. Except that what is important in a book is not the written letters, but the ideas that the letters give birth to in us, because they come from them. This is what differentiates a user manual describing the functioning of a machine, of a book, whose words refer not to a small number of unequivocal operations, but to ideas which are always the approximations of a total truth, transcending material existence. Where knowledge touches the spiritual. The greatest books, whose informative value is negligible, are precisely those whose movement leads us towards the absolute. Now, this absolute is an aim, not a synthesis of what is already known.

Also artificial intelligence is it badly named. To save ink and the worry of no longer having to write, we should rather speak ofclever artifice, an expression that emphasizes the materiality of the device rather than the intelligence from which it proceeds. A lifeless and therefore aimless materiality, capable only of optimizing a gigantic mass of information. The future cannot be an improved past, and intelligence the indefinite rehashing of what it has itself produced. Everything remains to be written.