There is a moment, when studying a musical instrument, when the body finally manages a movement that once seemed impossible.
A note finds its pitch, or a musical phrase acquires assurance and continuity. The listener perceives only the result.
All genuine learning shares this same nature, because learning means progressively changing how we observe, understand and interpret reality, rebuilding the learner from within rather than merely adding information.
For centuries, this dimension of knowledge remained hidden in practice: learning a language, understanding a text, solving a problem, acquiring a skill required a sequence of steps involving practice and engagement with error. The answer took shape at the end of a journey, and that journey was itself already an essential part of learning.
Today, this relationship is changing. Generative AI introduces a new possibility into the history of knowledge: obtaining a fully developed answer without having gone through the cognitive process. It is an extraordinary opportunity, capable of broadening access to knowledge and supporting creativity, research and productivity.
But we must ask ourselves what value the journey retains when the result is immediately available.
Knowledge becomes part of the person
Possessing information is different from being transformed by that knowledge. An experienced musician does not consciously think about every movement of the hand; a reader does not read every single letter before understanding a sentence; and the same holds for the mathematician, who does not reconstruct the entire logical system each time.
When studying the violin or another musical instrument, this principle reveals itself with an almost physical clarity. I remember well that stage in my students when the body is still searching for the movement, before the movement finds its own natural ease: a young musician learns to hear a minute difference in intonation, to anticipate a difficulty before it arises, to build an ever more precise relationship between intention and sound.
Musical competence is what makes that performance possible: the silent grammar that no listener perceives and every performer knows.
Error as a site of knowledge
Our relationship with error is essential to learning. In a culture of performance, error is often treated as a failing to be corrected quickly, an obstacle between the person and the desired result. Yet error serves an essential cognitive function: it makes visible the distance between what we know and what we have yet to learn.
When a student gets an exercise wrong, or when a researcher formulates a hypothesis that is not borne out, that moment yields valuable information about how thought works.
Writing to a young poet, Rilke urged patience towards everything that remains unresolved in the heart, learning to love the questions themselves.
Every teacher has observed that stage when a student seems stuck on a concept, until something suddenly changes: an understanding that appears instantaneous but is in fact the result of a long and often invisible process, made up of attempts, different explanations, questions and personal reworkings.
When the answer precedes thought
Artificial intelligence makes this mechanism more complex, because it takes the human being’s place in the space between question and answer. For centuries, the main difficulty was finding information to process and transform into knowledge. Today, the difficulty is different: distinguishing an available answer from knowledge genuinely acquired.
This distinction has a particular bearing on education. A student can use an artificial intelligence tool to understand a subject better, turning that technology into an extension of their own thinking.
Yet the same technology can replace the journey needed to build that thinking, especially when it intervenes before the fundamental skills required to assess, interpret and question what is produced have matured. A calculator helps those who already understand the meaning of a mathematical problem, while remaining incapable of building that understanding on its own. A satnav indicates a route, still leaving those who follow it with the task of developing a representation of the space through which they move.
Education in the age of AI
The educational challenge of the coming years will be to understand which formative experiences must continue to be lived directly, because they contribute to the making of the person.
Today, an answer can be generated in a few seconds; a mind capable of understanding it still takes shape only through the experience of time.