I call digital metacognition the capacity to observe oneself in one's technological uses. The aim is not first to pass moral judgement on those uses, but to describe them with lucidity: what happens to my attention after twenty minutes of conversation with an AI? What happens to my sleep when I move from one screen to another until late at night? What happens to my judgement when a machine formulates too quickly what I have not yet had time to work out for myself?
The word matters because it shifts the question. We talk too often about technical competence, as if it were enough to learn how to use a tool well. But the contemporary problem goes deeper: many digital uses are efficient while also being harmful. They allow us to move fast, but leave us more scattered. They give the impression of better understanding, while sometimes short-circuiting the interior work of formulation, hesitation and revision.
Observing before judging
This competence begins with a gesture of observation. One must learn to watch oneself act with screens and with AI as one would observe an experiment. Which moments leave me more lucid? Which uses leave me more nervous, more dependent, more impatient? At what point does a tool help me clarify my thought, and at what point does it begin to replace it?
This matters especially for conversational AI. Its real value begins when it helps us frame a problem, compare arguments, reformulate an idea, ask for an objection, open a line of inquiry. Its danger begins when it becomes the place where we delegate the work of searching, sorting, weighing and deciding. The line between the two is not always obvious. It is precisely the work of digital metacognition to make it perceptible.
An educational competence
Digital metacognition is a central educational competence. Autonomous subjects are not formed by transmitting only safety rules or lists of prohibitions. Students also need tools to name what they experience: diffuse fatigue, impatience, loss of continuity, the growing difficulty of holding a thought over time.
In a classroom, this can take simple forms: comparing a piece of work written with and without assistance; describing what changes in the rhythm of thought; keeping a brief journal of uses; distinguishing what belongs to help, comfort or substitution. These exercises are not peripheral. They are the ground on which a more demanding relationship to tools can gradually be built. A student who learns to observe the effects of assistance on their own work is already beginning to reclaim authority over their thought process.
From lucidity to autonomy
Digital metacognition is the first lever of responsible digital autonomy. A subject who knows how to observe themselves in action can adjust their uses, set limits, choose their mediations and refuse certain conveniences.
It is also inseparable from an ethics. A useful AI is not merely a capable AI. It is an AI whose use can be integrated into a human life without destroying attention, interiority, judgement or the quality of relationships. Digital metacognition is the condition under which such an integration remains conscious rather than suffered.
It connects directly to digital anomie: one of the responses to collective disorientation passes precisely through this capacity to name effects, perceive what is at stake and resume a deliberate rather than automatic relation to tools. Without this individual and shared lucidity, even the best-designed charters and frameworks remain external impositions. With it, they become a lived grammar.
Frequently asked questions
What is digital metacognition? It is the capacity to observe one's own uses of technology with lucidity: to notice what they do to attention, effort, judgement and the relation to reality, in order to preserve a deliberate rather than automatic relationship with tools.
Why is it a central educational competence? Because autonomous subjects are not formed by rules and prohibitions alone. They need tools to perceive, name and evaluate the effects of digital uses on their own formation.
What does it look like in practice at school? Comparing work written with and without assistance, describing changes in the rhythm of thought, distinguishing help from substitution, and developing shared reference points to evaluate uses of AI rather than simply submit to them.