Laurence Plazenet is a professor at the University of Clermont-Auvergne, director of the Blaise-Pascal International Center, president of the Society of Friends of Port-Royal and honorary member of the Institut universitaire de France. She won the European Union Prize for Literature in 2012.
After the Football World Cup and the funeral of the Queen of England, ChatGPT – a more pleasant subject than the resurgence of war in Europe or the consideration of our death – has occupied editorialists, fine minds and curious people these last few days.
The discovery of artificial intelligence software has stunned neophytes.
Enthusiastic about the answers or the presentations he provided them, they were amused that the teachers would now have a lot to do to correct the impeccable copies of their pupils, or their students, who would have requested it.
To which is added that ChatGPT converses with lonely souls, takes the microphone, kits the verse, writes books and treatises.
Miracle.
Artificial Intelligence: Should we fear the ChatGPT revolution?
Admittedly, the glare ebbs a little.
ChatGPT lets itself be deceived by false information.
His creations are mediocre and often faulty tracings, if he decides, for example, to experiment with prosody.
Nourished by the vast literature available on the web, literature that he chews over according to statistical and probabilistic algorithms capable of reprogramming themselves, it is said, according to procedures that their initial programmers cease to master, he flounders in the ready- thinking.
So there are few ideological proprieties to which he does not pay the expected homage.
These are venial sins that it would be possible to correct upstream.
They don't mark where the ChatGPT illusion really is, or what it reveals about a world that doesn't outright denounce it in a hearty laugh.
Not for a moment does this enlightened robot think.
A fortiori in the country of Descartes, ChatGPT, is, as such, nothing.
Laurence Plazenet
ChatGPT produces summaries, reading notes, general presentations: this enlightened robot does not think for a moment.
A fortiori in the country of Descartes, ChatGPT, is, as such, nothing.
Because this program never questions, never asks a question.
However, it is not knowledge, but only questions that we strive to reduce, refine, sharpen, resolve, using knowledge transmitted by tradition, knowledge acquired over a long period of time, patiently matured, sometimes finally dominated at the cost of years of study and reflection, only after which it becomes possible to analyze them and submit them effectively to criticism, to try, in all humility, knowing that, already, our own revisers are probably coming to be born, to overcome our heritage and to make it progress.
Knowledge is a long chain of masters and disciples, of falling outs, of outbursts, of heated debates, of errors, of fervent transmissions and of bifurcations that were unpredictable, or even unthinkable.
ChatGPT, he is harping.
He responds to “course questions” with mediocre syntheses, without momentum, without gleam, without surprise, without genius.
ChatGPT drools over its acquired knowledge, clinging to it like a hermit crab to its shell or a mussel to its rock.
Poor ChatGPT that over a million pranksters try to stick around to see when it's going to make a fool of itself (which does happen).
ChatGPT, he is harping.
He responds to “course questions” with mediocre syntheses, without momentum, without gleam, without surprise, without genius.
ChatGPT drools over its acquired knowledge, clinging to it like a hermit crab to its shell or a mussel to its rock.
Poor ChatGPT that over a million pranksters try to stick around to see when it's going to make a fool of itself (which does happen).
ChatGPT, he is harping.
He responds to “course questions” with mediocre syntheses, without momentum, without gleam, without surprise, without genius.
ChatGPT drools over its acquired knowledge, clinging to it like a hermit crab to its shell or a mussel to its rock.
Poor ChatGPT that over a million pranksters try to stick around to see when it's going to make a fool of itself (which does happen).
ChatGPT, maker of school copies?
Who can have a rather stupid representation of what one asks of a pupil or a student to write such nonsense?
A dissertation is an agreed rhetorical exercise, intended to verify the acquisition of knowledge, the mastery of linguistic tools and the intelligence of a subject (i.e. the good intellection of the subject to be treated and the aptitude mental state of the individual to whom he is subjected to illustrate it and to perceive its weaknesses, which will allow him to advance from there an original and more relevant personal proposal).
This algorithm has none of the resources to do this and it fails pathetically.
No conscientious student, on the other hand,
would like to entrust him with a duty where it is up to him to test his own aptitudes, to develop them, to harden them in the manner of any athlete.
As for the lazy ones, rest assured: their average level is so destitute that a copy stamped ChatGPT has no chance of deceiving any of their teachers.
ChatGPT, maker of theses or essays?
Neither have anything in common with course questions or preparatory exercises that are worth memorizing data, which is the responsibility of ChatGPT.
They aim to increase the scope of our knowledge or our reflection, and therefore are based first on a questioning.
With this program, no valid questioning: we will only have chatter.
Theses and essays are questions, pure and jubilant pas de deux, de trois, de four, de five, de six... They are dances, mental projections that unfold like the body of the dancer on the stage, in freedom, expressions of an intellectual and sensitive singularity forged in the same suffering, and the same joy, as those of the dancer, in order to tame the
Let's face it, ChatGPT is perfectly silly.
He does not have the imagination to question, that is to say he does not think.
He has nothing to teach us.
Laurence Plazenet
Do we ask ChatGPT about Pascal and prophetism or the prophets?
He eludes a fascinating and almost new subject by a banality.
“
Nil novi sub sole
”
,
he answers you in his learned and phraseful way.
Does one request from one's intelligence a subject of communication for a colloquium?
ChatGPT has no idea.
He replies that everything is said.
Go your way.
Nothing to see.
Nothing to think about.
Nothing to ponder.
Subway, work, sleep, friends.
Or Candy Crush.
Ask ChatGPT for a summary of China-Japan relations from the Ming reign to the end of World War II?
ChatGPT shudders.
He answers.
The paragraphs line up.
But what to do with a hodgepodge which, not once, indicates its sources?
Who never does the chronology according to which a problem has been solved, nor that of the solutions envisaged?
Who demands, therefore, that we rely blindly on his letter?
Shall I repeat his presentation like an imbecile robot who trusts desperately to a kind of catechism which is neither to be discussed nor to be understood, that is to say to be grasped in a clear awareness of its origin, of its environment? elaboration and influences, even determinisms, that one and the
another weigh on him?
Ask ChatGPT for an original and elegant mathematical demonstration?
He knows theorems.
On the other hand, he will not be able to imagine the inspired subtlety which suggests, for the first time, to pass through such intermediate demonstrations, and to use familiar theorems in what apparently constitutes a digression "on every point which relates to the end", to hit the target.
Neither student, nor teacher, nor researcher, will ChatGPT be a writer?
But, again, he does not imagine.
He does not create without outside solicitation.
You have to place an order with him, an order as precise as possible.
The spirit of the times and the inflation of “creative writing” courses make him a child of his century, no doubt.
What, however, is its added value?
Its necessity?
ChatGPT does not even pastiche: it needs a bit of irony or humor, distance, when our software sticks to its doxa.
He is therefore content to imitate.
Flatly.
Benevolently.
Without malice.
Forger without ambition or banter.
Let's face it, ChatGPT is perfectly silly.
He does not have the imagination to question, that is to say he does not think.
He has nothing to teach us.
Nothing to tell you.
Wouldn't he be a machine (a
money pump
?), that he would attract our compassion: poor little organon, enslaved and blind.
But mistaking ChatGPT, and devoting so much attention to it?
How is this possible after Homer, Socrates, Leonardo da Vinci, Shakespeare, Einstein and Pascal?
Is it abysmal stupid or, by the effect of a reversal of the pros and cons, the occasion on the contrary to draw one or two lessons from it?
Isn't it time to wake up, to rebel against the great moralizing consensus instilled in AI, to open a Plato and go and think a little, with Socrates, on the edge of Ilissos?
Laurence Plazenet
To misunderstand ChatGPT, in reality, is to misunderstand intelligence.
About what we think it is, what it resides in, how it grows and thrives.
This is to misunderstand what we believe education is for our children.
Do we think that it is destined to make good little ChatGPTs out of them, disgusting sterilely in one language or the other the same eternal axioms?
Do we want to manufacture “artificial intelligences”
,
or men, fallible and yet without equal (“thinking reeds”, it was said around 1660)?
Puppets or free spirits, dancing spirits, mobile and flexible, so that they will catch a glimpse of the unexplored and invisible path where the future of thought, even of science, lies?
ChatGPT?
The wind.
A wisp of straw.
But a speck in our eye.
How can he abuse us so much?
Where did we come from by dint of compromises and intellectual renunciations to consider for a moment that this nonsense will replace the least of us?
How can we be so satisfied with manuals "for dummies"
,
ready-to-repeat, opinions, AFP dispatches repeated identically from news site to news site, so as not to immediately sniff out the scam and, in disbelief, laugh at us in front of this miserable joke?
This is precisely where the shoe pinches.
Thousand-and-one activities required of one or the other, thousand-and-one articles written today, thousand-and-one of these digests set up in bibles of modern times, thousands of false debates and cathodic disputes of which one waters us, correspond to ChatGPTesque statements.
Yes, accepting to submit to this intellectual indigence and feeding ourselves on it, we are worth no more than the Californian robot and its insipid logorrhea.
We become replaceable.
Isn't it time to wake up, listen to our researchers and scholars, rebel against the big moralizing consensus instilled in AI neurons, I mean open a Plato and go think a little, with Socrates, at the edge of Ilissos?