Of course, AI would write a better philosophy essay than yours. But it wouldn't be worth anything.

Published on June 17, 2025 at 6:20 p.m. , updated on June 17, 2025 at 6:20 p.m.

Google News

Reading time: 2 min.

Philosopher Wilhelm von der Weid explains why algorithms, however erudite they may be, cannot produce a thought.

This article is an op-ed, written by an author outside the newspaper and whose point of view does not reflect the editorial staff's views.

The philosophy baccalaureate has just ended, and it is no coincidence that this year's two essay topics question a new reality, partly shaped by artificial intelligence: "Does our future depend on technology?" and "Is the truth always convincing?" . This is because we need smart minds more than ever, rather than full-fledged ChatGPTs.

Many students are worried about their ability to meet the expectations of the philosophy essay, a test that is both codified and mysterious, solemn and pointless. And the candidates wonder: what would ChatGPT have done in my place? Because artificial intelligence is now not only capable of writing a perfect philosophy essay, but also of correcting it. What's the point of thinking for yourself if the robot thinks better, and everything, including itself?

Also read

Investigation: Who is John Rawls, the American philosopher betrayed by his French fans and the subject of the 2025 baccalaureate exam?

A question that is in reality as old as a philosophy with an ambiguous status, between knowledge that one simply needs to learn – in the same way as mathematical theorems, historical events and scientific discoveries – and a personal opinion with limited scope. “Useless and uncertain” , as Blaise Pascal ironically summed up.

But if ChatGPT can automatically develop a thought based on all available knowledge, it means that we can have a correct thought, without having to think for ourselves. A universal thought without subject, without bias, without holes, would result, through the detours of a totalizing, fine and structuring algorithm, in a perfect dissertation that would only be a matter of copying.

Problem solved.

Except that philosophy is not just a baccalaureate exam: it is also a subjective test, which consists of becoming aware of the limits of one's knowledge, and not only of its quantitative boundaries, to use the Kantian distinction. For however extensive my knowledge may be, it will never reach the world in its ultimate reality, the source of morality, the mystery of religion, the meaning of life. Now philosophy is precisely the test of my finitude, of my unconscious, of my mortality, that is to say of problems that have no solution.

Also read

However, let's count on ChatGPT to make a solution: The solution is that we must accept that there is no solution! It is not a solution, it is the human condition. ChatGPT can synthesize everything that has already been said about these problems in a perfect way, it will make a beautiful summary note, not a human thought, which is only thought by being experienced .

But then, how can we evaluate an inner trial by a school test? This is the whole point of the philosophy class, where we practice thinking together, in dialogue, without presuppositions. I have been saying this in every tone to my students for twenty years: no quotes, no theory, no lectures, but only a thought that progresses through your experience, your questions, your aspirations. A thought that takes seriously the problems, which are your problems, your illness, the father who abandoned you , the truth you avoid, the love you doubt, the morality you are incapable of...

It is through this life that questions itself, rather than through "topos" [commonplaces] about authors, notions, and benchmarks, that the best copies are written. Copies that are oh so imperfect, but alive, which show a autonomous progression, which is the true goal of school. Because a society does not want perfect people, robotic superhumans, but people who know how to judge, judge that they do not know, judge what we do with knowledge.

But above all, judging what knowledge is, at a time when AI threatens to submerge the planet with a flood of automated articles, deepfakes , self-perpetuating loops on social networks, which irreversibly mix the true with the false. ChatGPT does not know how to judge.