Why human limits are a source of value in the age of AI?
Machines do not get tired, do not forget, do not die. That is precisely why they do not make any decision that truly matters. Human finiteness is not a flaw to be compensated for: it is the source of the value of our choices.
A limited amount of time changes the weight of our choices
A being that lives forever can afford to postpone everything; a finite being cannot. Every hour you dedicate to a project is an hour taken away from everything else in your life: that is why your presence in a project means something. Machines can generate a thousand plans in a minute precisely because no plan costs them anything. The scarcity of human time is what turns a choice into a commitment.
Machines have nothing at stake
An AI model does not risk its reputation, does not feel shame over a mistake, does not celebrate a result. It can simulate these states, not live them. When an engineer signs off on a project, they put something of themselves on the line; when an AI produces an output, nothing is at stake. It is this asymmetry, not intelligence, that represents the real difference: responsibility requires someone who can lose something.
The body as an instrument of knowledge
Humans also gain knowledge through the body: the exhaustion on a construction site, the discomfort in a meeting that is going badly, the hesitation of a client who says yes in words. This embodied knowledge guides decisions that no data can capture. An AI can read the meeting minutes; it cannot feel that the temperature in the room has changed.
Finiteness generates care
We take care of what can break, die, end. Concern for the planet comes from exactly this: from recognising that resources are finite, that ecosystems can collapse, that generations pass. An intelligence without finitude can calculate sustainability; only those who are finite can feel it as urgency. This is why impact work remains deeply human even when it uses the most powerful tools.
The risk: outsourcing what defines us
The temptation of the AI era is not that machines will replace us — it's that we hand over to them, piece by piece, the activities that make us capable: evaluating, choosing, remembering, writing the first thought instead of the prompt. Every delegation is legitimate as long as it remains reversible: as long as you still know how to do, and above all judge, what you have delegated. When delegation becomes dependency, the human limit stops being a compass and becomes a cage.
The compass, in practice
Before delegating a task to AI, three questions: can I still judge the quality of the result? Can I explain why this choice and not another? If tomorrow the tool disappeared, could I redo it well enough? If the answers are yes, the delegation is empowering you. If they are no, it is hollowing you out. Throughout the rest of the programme we will see how to keep the yes on judgement, responsibility and ethics — the three things that are never delegated.
Frequently asked questions
Is human finitude a disadvantage compared to AI?
No: it is the source of the value of choices. A limited time makes every commitment meaningful, and the possibility of losing something is the foundation of responsibility — which machines cannot have.
What does reversible delegation to AI mean?
Delegating a task while remaining capable of judging its result and, if needed, redoing it. Delegation empowers as long as it stays reversible; when it becomes dependency, it weakens.
Why is caring for the planet linked to finitude?
We care for what can end. Recognising finitude — of resources, ecosystems, time — is what transforms sustainability from a calculation into a felt urgency.
← The full course · Next pill: Judgement: deciding without certainty →