règle : θ = n × 137,508° · r = c√n · graine : cognitive-hybridisation

Concept entry · Anthropology of technology | Cognitive science · original concept

Cognitive hybridisation

A fundamental human faculty for extending and transforming oneself by integrating tools, prostheses and technological extensions into one's cognitive processes — AI prolongs a hybridisation that is constitutive of the species, it does not inaugurate it.

You hand a route over to your GPS and you sense, obscurely, that you would no longer know how to find it again without it. The thought is unpleasant: something is wasting away, a capacity is slipping from you, you are becoming dependent. The same reasoning is applied to AI — it will make us incapable of thinking, of writing, of deciding for ourselves. The worry is sincere, and it rests on a very old image: that of a “pure”, whole human being, who ought to be protected from the contamination of tools. The trouble is that this human being has never existed.

The tree and its grafts

Cognitive hybridisation is the name for what we have always done: extending ourselves through our tools, and transforming ourselves in the process. A tree gives the right picture. The trunk has stayed the same, but each graft — knapped stone, fire, writing, the printing press, the computer — has widened its canopy without denaturing it. We are not a pure trunk threatened by foreign grafts; we are the grafted tree, and have been for so long that the graft can no longer be told apart from the wood.

This shift changes everything. As long as we believe in the pure human, every new tool is a loss to be feared. As soon as we see the tree, AI is no longer a rupture but the latest graft in a very long lineage — neither a parasite nor a miracle: one more graft, to be made a success or a failure like the others.

What moves when we delegate

There remains the original fear, which must not be brushed aside: do we not lose skills when we delegate? We do. It is true, and it must be said plainly. Mental arithmetic dulls with the calculator, the sense of direction with the GPS. But let us look at what really happens: the skill does not vanish into the void, it moves. The calculator did not make mathematicians more stupid; it allowed them to tackle problems that no calculation by hand could reach.

No one regrets no longer being able to run fifty kilometres to carry a message now that the telegraph exists. The question is never “what do I lose?” on its own — it is “what do I lose, and what does that free up?”. A successful graft redistributes; it does not merely subtract.

It will be objected that this reasoning permits anything: if hybridisation is our nature, then any hybridisation would be good, and there would be nothing left to criticise. That would be a lazy reading, and a dangerous one.

For the concept describes a faculty, not a duty. To say that we can hybridise ourselves says nothing about what we must accept. Some grafts emancipate — they make us more capable of thinking, while keeping our hand on the wheel. Others alienate — they make us dependent on a system we no longer understand and no longer control. Telling the two apart is precisely the work the concept makes possible: three questions are often enough — do I still understand what the tool is doing? can I do without it if need be? am I gaining a capacity, or only convenience?

One last reservation, which is no small matter: if hybridisation is an advantage, then unequal access to good tools becomes a cognitive inequality. What literacy was to writing, lucid access to AI may well be to augmented thought. The concept does not resolve this question — it makes it urgent.

What it helps us think

This concept does not say that AI is good for us. It simply takes a false alarm out of play — that of the “loss of humanity” — to make room for the real question, which is more demanding: which grafts do we want, on what conditions, and who decides? It is the same gesture, at the scale of the individual, that ecosystemic intelligence applies to the collective: intelligence is not a property one possesses and might lose, it is a relation one cultivates.

There remains the question this note leaves open, and that must be carried away: of your own hybridisations — the tools you already think with, without thinking about them — which ones have made you more capable, and which have only made you dependent? The answer is not to be read in the tool. It is to be read in what would be left to you if it disappeared.

What this concept is not

It is not a replacement. Replacement substitutes the machine for the human — the machine alone does the work, and the human disappears from the equation. Hybridisation integrates the tool into a practice that remains human: the result is the human plus the technology, not the technology in place of the human. A doctor supported by a diagnostic AI makes a better diagnosis; an AI that diagnoses on its own is something else, and no longer falls under hybridisation.
It is not a loss of identity. The most common objection equates extending oneself through a tool with a dissolution of the self. The opposite is true: human identity is not a fixed essence that a tool would erode, but a process that transforms itself as it equips itself. We did not become less human by learning to write; we became human otherwise.
It is not a prescription. To say that hybridisation is our age-old condition does not mean that every hybridisation is good. The concept describes a faculty, not a duty: some extensions emancipate, others alienate, and telling the two apart remains a work — the very one the concept makes possible.

Examples

Education: training for conscious hybridisation

What is at stake: learning to hybridise one's capacities effectively with AI tools

Skills to develop:

Discernment: when to delegate to AI, when to keep human control?
Meta-cognition: understanding how AI transforms our thought processes
An ethics of hybridisation: which hybridisations are desirable?
Hybrid creativity: making the most of human–AI synergies

Practical example:

Writing course + AI: learning to use AI for drafts and structuring, but keeping the final judgement human
Not banning AI (a sterile resistance) but training for conscious hybridisation

Work: redefining roles (doctor, lawyer, designer)

Principle: reconceiving occupations as human–AI hybridisations rather than competition

Examples:

Occupation	AI tasks	Human tasks	Hybridisation
Doctor	Analysis of medical images, diagnostic suggestions	Dialogue with the patient, final decision, empathy	AI + doctor = a more reliable diagnosis
Lawyer	Case-law research, drafting of standard contracts	Strategy, advocacy, client counsel	AI + lawyer = a more efficient service
Designer	Generation of visual variations	Selection, refinement, creative vision	AI + designer = accelerated exploration

Benefit: refocusing humans on high-value tasks (relationship, judgement, creativity), delegating repetitive tasks to AI.

Mental health: supporting adaptation

Observation: some people experience hybridisation with anxiety (fear of dependence, loss of skills)

Therapeutic approach:

Step 1 – Validation:

Acknowledging the legitimacy of anxiety in the face of change
Normalising the discomfort (every transition = a temporary imbalance)

Step 2 – Psychoeducation:

Explaining the history of hybridisation (always done)
Showing that dependence on GPS is not different in kind from dependence on paper maps (same process, new medium)

Step 3 – Restructuring:

Replacing “I am losing my capacities” with “I am transforming my capacities”
Replacing “I am becoming dependent” with “I am extending my possibilities”

Step 4 – Experimentation:

Trying gradual hybridisations (AI for simple tasks first)
Observing concrete benefits (time saved, improved quality)

Technological design: designing for hybridisation

Principle: creating AI tools that facilitate harmonious hybridisation, not replacement

Design criteria:

Transparency: the user understands what the AI is doing
Control: the user keeps the power of final decision
Complementarity: the AI compensates for human weaknesses, the human compensates for the AI's weaknesses
Mutual learning: the AI adapts to the user, the user learns from the AI

Example: assisted-writing tools (vs fully automatic writing)

AI suggests, human chooses
AI structures, human refines
Result: an authentically human text, enriched by AI

Other perspectives

Objection — if hybridisation is “natural”, does that make every hybridisation good? Reply: no. The concept describes a faculty, not a moral prescription. Some hybridisations alienate: pathological dependence, loss of critical judgement. Three criteria distinguish a healthy hybridisation — does it genuinely increase capacities? does it preserve autonomy and judgement? does it remain reversible?
Objection — does hybridisation with AI not atrophy our skills? Reply: some do dull, it is true — mental arithmetic with the calculator, the sense of direction with the GPS. But they do not vanish into the void: they redeploy. The calculator did not make mathematicians more stupid, it opened up to them problems beyond the reach of calculation by hand. Hybridisation redistributes skills, it does not merely subtract them.
Objection — does hybridisation not open up new inequalities? Reply: it does, and it is a major issue. If access to good tools becomes a cognitive advantage, its unequal distribution opens a rift. What literacy was to writing, lucid access to AI may well be to augmented thought: the concept does not resolve this question, it makes it urgent.

The concept in detail

Physical Tools (Prehistory — 2.5 million years ago)

First stone tools:

Extension of manual capacities (cutting, grinding, shaping)
Transformation of the relation to the world (hunting, habitat, food)
Already a form of hybridisation: human + tool ≠ human alone

Fire (~400,000 years ago):

Extension of metabolic capacities (digestion of cooked food)
Social transformation (gathering around the hearth)
Deep hybridisation: fire is not “external” to the human, it redefines the species

Language and Writing (Cognitive Revolution)

Spoken language (~70,000 years ago):

Radical extension of cognitive capacities
Transmission of knowledge across generations
Symbolic hybridisation: language reconfigures human thought

Writing (~5,000 years ago):

Externalisation of memory (no need to retain everything)
Accumulation of knowledge beyond the span of an individual life
Cognitive transformation: thought organises itself differently with writing (Ong 1982)

Printing press (1450):

Democratisation of access to knowledge
Standardisation of knowledge
Collective cognitive revolution: a new form of social intelligence

Machines and Automation (Industrial Revolution)

Industrial machines (18th–19th centuries):

Extension of physical capacities (strength, speed, precision)
Transformation of work and social organisation
Productive hybridisation: human + machine > human alone

Electricity and telecommunications (19th–20th centuries):

Extension of sensory capacities (telegraph, telephone, radio)
Reduction of temporal/spatial distance
Communicational hybridisation: a new relation to space and time

Computing and the Digital (Digital Revolution)

Computers (20th century):

Extension of computational capacities
Automation of complex calculations
Analytical hybridisation: delegation of repetitive cognitive tasks

The Internet (late 20th century):

Externalisation of collective memory (Wikipedia, databases)
Instant access to planetary knowledge
Informational hybridisation: memory becomes external and distributed

Smartphones (21st century):

Integration of multiple extensions (GPS, calculator, camera, communication)
Permanent connection to the global network
Ubiquitous hybridisation: cognitive extension becomes portable and ubiquitous

Artificial Intelligence (Revolution Under Way)

Conversational, generative and analytical AI (2020s):

Extension of creative capacities (generation of text, images, code)
Extension of analytical capacities (pattern recognition, prediction)
Extension of decision-making capacities (diagnostic support, strategy)

Specificity of AI:

Not a qualitative rupture: continuity with the externalisation of memory (the Internet) and the automation of calculation (computers)
A quantitative difference: an unprecedented level of sophistication + versatility + accessibility
A new evolutionary stage: extension into domains previously “exclusively human” (creativity, dialogue, judgement)