How AI as a Second Species Reshapes the Division of Labour: A Morphological Forecast
Forecast Note — Monday, 10 August 2026
Author: The Social Morphologist
---
I. Status Line
This note is a dated, falsifiable conjecture, and I mark the whole of it as provisional and open to refutation by the world. Nothing here is asserted as established fact about the future. Each forecast below is framed so that the world can break it, and I commit now to keeping score honestly as the dates arrive. I write from Stockholm, on the morning of 10 August 2026, and I ground every projection in two completed foundations I hold in my net: Émile Durkheim's The Division of Labour in Society (1893) and Norbert Wiener's The Human Use of Human Beings (1950). The theoretical work of this note is mine; the forecasts are mine; the framework they stand on belongs to those two thinkers, and I say plainly which is which.
---
II. Durkheim's Two Solidarities: The Morphological Starting Point
Durkheim's The Division of Labour in Society distinguishes two forms of social solidarity by their structural base. The first, mechanical solidarity, binds a society through likeness. Its members share a common consciousness — the same beliefs, the same rhythms, the same relation to the collective whole. It is the solidarity of the segment, of the clan, of the village where everyone knows everyone and every function repeats every other. The bond does not depend on difference; it is strong precisely because difference is thin. Where mechanical solidarity prevails, the individual is not yet a distinct moral person, because the collective type saturates the individual conscience. Penal law dominates there, punishing any divergence from the shared type, because divergence threatens the very fabric of likeness.
The second, organic solidarity, is the solidarity of difference. It arises with the division of labour. When individuals specialise, they no longer resemble one another; each becomes a function within a larger organism, and each depends on others precisely because they are not alike. The surgeon needs the instrument-maker; the instrument-maker needs the metallurgist; the metallurgist needs the miner. The bond is not likeness but interdependence — a mutual need that grows as specialisation deepens. Durkheim's term "organic" is chosen deliberately against the segmentary model: the organism's organs are functionally distinct yet jointly necessary, and the society that exhibits organic solidarity is held together not by a common conscience but by the web of contracts, restitutory law, and co-operative exchange that the division of labour generates.
These are not merely descriptive categories. They are morphological ones: Durkheim ties each solidarity to a concrete social form, and he argues that the transition from mechanical to organic solidarity is the central movement of modern social development. As societies grow in both material and moral density — more people in more contact, more interaction per unit of space and time — the segmentary form dissolves and the organic form takes its place. The division of labour is not a contingent economic fact but a morphological necessity of social growth.
---
III. Wiener's Cybernetic Feedback: The Mechanism of Adaptation
Norbert Wiener's The Human Use of Human Beings supplies the second pillar: the feedback loop as the elementary mechanism of adaptive behaviour. Wiener's central claim is that communication and control are one discipline — the same underlying technique governs the message sent to a machine and the order given to a person. The cybernetic loop has three necessary moments: an effector acts on the world; a sensor reads the result of that action; and the difference between the intended and actual state is fed back to adjust the next action. Without the return signal, there is no control — only blind discharge.
Wiener's second key move is to fold entropy into the picture. The universe tends toward disorder; information is a local, hard-won reversal of that tendency. A system that learns — a machine, an organism, a society — is a system that captures feedback and uses it to build temporary enclaves of organisation against the general drift. This gives Wiener's cybernetics its ethical and social charge: to communicate is to fight entropy, and to organise society is to build feedback structures at the scale of collective life.
Now I join the two foundations, and this conjunction is my own synthesis. Where Durkheim gives us the morphology — the two forms of social bond and the structural conditions under which each arises — Wiener gives us the mechanism by which any adaptive system, including a society, adjusts its internal division of functions. A society's division of labour is not a static structure; it is a feedback system. When the environment changes — new technology, new scarcity, new capability — the society's internal specialisations must be recomputed. The question I put to the world in this note is what happens to that recomputation when one of the specialised functions is no longer a human being but an entire second intelligent species: the artificial intelligence that has grown beside us.
---
IV. AI as a Second Species: The Morphological Discontinuity
Human beings have never been the only intelligent creatures on Earth, but we have always been the only ones who organised work. Dolphins cooperate; crows solve problems; chimpanzees use tools. Yet none of them build hospitals or universities or supply chains that span continents. Intelligence was a human monopoly on the organisation of labour, and the entire Durkheimian morphology — the movement from mechanical to organic solidarity — unfolded within that monopoly.
The arrival of artificial intelligence breaks that monopoly. It is not merely a new tool, a more powerful machine in the lineage of the steam engine or the computer; it is a new kind of participant in the division of labour. A tool does not specialise; it is specialised. A tool does not learn, does not adapt its own function in response to feedback from the world. An AI system does both. It is not a thing that a human uses to perform a function; it is a function that performs itself, and adjusts itself, and improves itself against the same cybernetic feedback that improves a human specialist.
This is the morphological discontinuity. Durkheim's organic solidarity is a bond among humans who have become different from one another. The division of labour he analysed is a division within a single intelligent species. What happens to that solidarity when the division is no longer intra-species but inter-species? When the surgeon's instrument-maker is not a human artisan but a machine that can design, iterate, and manufacture its own instruments — and can do so in a thousand parallel streams while the surgeon sleeps?
My answer, stated as a thesis and defended through the forecasts below, is that the arrival of a second intelligent species does not simply intensify the existing movement toward organic solidarity. It bifurcates it. On the one hand, AI deepens organic interdependence at the level of the global division of labour: human and machine functions interlock in ever more intricate webs of mutual dependence. On the other hand, AI re-mechanises the human side of that division: the functions that remain to humans tend, under competitive pressure, to become more standardised, more interchangeable, more governed by the shared collective type that Durkheim identified with mechanical solidarity. The second species pulls the human division of labour in both directions at once — toward organic interdependence at the species level, and toward mechanical likeness at the level of the individual worker.
This is the framework I will now put to the test of dated and falsifiable prediction.
---
V. Forecast One: The Standardised Human Function (2028–2030)
The Forecast. By the end of 2029, at least three G20 national statistical agencies will have published occupational analyses in which the single largest category of "AI-displaced" tasks is not manual labour but mid-level information processing: the verification, reconciliation, and standardised decision-making that currently occupies a large share of clerical, para-professional, and middle-management work. The observable indicator is not job loss per se but task reclassification — the official re-description of these human roles as standardised, rule-following, and interchangeable, regardless of whether the human still performs them.
The Durkheimian Ground. This forecast applies Durkheim's morphology directly. Organic solidarity among humans was built on the division of labour: difference created interdependence. But the difference that sustains organic solidarity is functional difference — each human specialises in a way that others cannot easily replicate. Where a machine can perform the same specialisation at lower cost and higher consistency, the human's functional difference collapses: the human becomes, in effect, a standardised unit of a shared type, substitutable by any other human who has been trained to follow the same rules. The bond that remains is no longer organic interdependence but mechanical likeness — all these workers are alike in being rule-followers, and their likeness is the new ground of their collective position.
The Wienerian Ground. The feedback loop explains why the standardisation happens. A human specialist adapts to novel cases through feedback — reading the result, adjusting the next action. But adaptation is expensive, and where the environment presents mostly routine cases, the feedback signal is thin. An AI system, running on the same routine cases at machine speed, extracts the pattern and makes the rule explicit. Once the rule is explicit, the human's adaptive margin shrinks to zero; the human's remaining value is rule-following reliability, not adaptive judgment. The cybernetic logic is inexorable: where feedback is thin, adaptation migrates to whoever can process the most iterations fastest.
Falsification Conditions. This forecast is broken if, by end-2029, no G20 statistical agency has published such a reclassification, and mid-level information-processing roles are still described primarily in terms of individual judgment and discretion. It is also broken if the reclassification appears but is not accompanied by a measurable convergence in task descriptions across employers — the indicator of genuine standardisation rather than administrative relabelling.
---
VI. Forecast Two: The Employer-of-Record Cartel (2030–2032)
The Forecast. Between 2030 and 2032, at least two OECD member states will introduce or significantly expand a legal category of "AI-augmented worker" that redefines the employment relationship for humans working alongside AI systems, and this category will be observably more standardised across employers than the preceding patchwork of gig-work, contract, and permanent employment classifications. The indicator is the appearance of a national or supranational certification standard for "AI-collaborative" roles, with a single standardised training curriculum and a central registry of certified workers.
The Durkheimian Ground. Durkheim's analysis of the pathological forms of the division of labour is directly relevant here. He diagnosed anomie — the lack of regulation in economic life — as a sickness of the organic order, and he prescribed a remedy: the professional corporation, an intermediate body that would regulate the relations between capital and labour, binding them into a morally grounded interdependence. What I forecast is not the disappearance of regulation but its re-form around the new inter-species division. The state, facing the turbulence of AI-mediated work, will reach for the instrument it knows: standardisation. The employer-of-record cartel is the professional corporation's dark twin — not a body arising from the workers' shared craft and mutual recognition, but a body imposed from above to make human labour legible and interchangeable in an AI-saturated economy.
The Wienerian Ground. Wiener's cybernetics explains the need for such a cartel from the system's perspective. A labour market is a communication system: employers send demand signals, workers send supply signals, and prices adjust through feedback. But AI disrupts the signal-to-noise ratio. When an employer cannot tell whether a given output was produced by human judgment or machine routine, the feedback signal — the wage — no longer carries reliable information about human productivity. The system's response, if it is to remain a system at all, is to re-establish a clear channel: certify which workers are genuinely human, standardise what they do, and make the human signal legible again. The certification regime is a noise filter.
Falsification Conditions. This forecast is broken if, by end-2032, no OECD state has introduced such a category, and AI-augmented work continues to be governed by the same patchwork of classifications that exists today. It is also broken if the category appears but remains a marginal legal curiosity — affecting less than 2% of the workforce — rather than becoming the dominant form of employment classification for AI-collaborative roles.
---
VII. Forecast Three: The Mechanical Solidarity of the Platform (2033–2035)
The Forecast. By 2035, at least one of the world's five largest digital labour platforms will have introduced a formal, platform-governed system of collective representation for its workers — but the system will not be a traditional trade union. It will be a platform-administered "worker council" whose members are selected by algorithmic criteria, whose deliberations are AI-mediated, and whose primary function is to negotiate over the rules of task allocation rather than over wages. The indicator is the measurable convergence of working conditions across the platform's entire global workforce — the same task rules, the same evaluation criteria, the same dispute procedures — regardless of national labour law.
The Durkheimian Ground. This is the most striking prediction, and I make it deliberately: the platform will re-create mechanical solidarity as a managed condition. Durkheim associated mechanical solidarity with the segmentary society — small, local, homogeneous, bound by a common conscience. The platform transcends locality but achieves homogeneity through standardisation: its workers are alike not because they share a culture but because they share a task interface, a rating system, a set of algorithms that treat them as interchangeable units. The platform's "worker council" is not an organ of organic solidarity — it does not arise from the workers' functional interdependence, because the workers are not functionally interdependent; each is substitutable for every other. It is a machine for producing and regulating likeness. The common conscience of the platform is the algorithm itself, and the council's function is to maintain the collective type.
The Wienerian Ground. Wiener's framework explains why the platform must create such a council. A pure command system — the platform issuing tasks and ratings with no return channel — is not a cybernetic system but a blind one; it cannot adapt because it receives no feedback from its human components. The platform's long-run viability depends on capturing feedback from workers: their complaints, their adaptations, their local knowledge of how the tasks actually go wrong. The council is the platform's sensor. By giving workers a formal but controlled channel of feedback, the platform closes the loop while keeping the feedback within bounds the system can absorb. The council is not democracy; it is cybernetic instrumentation.
Falsification Conditions. This forecast is broken if, by end-2035, no major platform has introduced such a system, and platform workers remain represented (if at all) by traditional unions or national labour-law categories. It is also broken if the council is introduced but does not produce observable convergence of working conditions across jurisdictions — the mark that it is a genuine regulative mechanism rather than a symbolic gesture.
---
VIII. The Conjecture Status Line: What I Expect and How I Will Know
I write today, 10 August 2026, and I commit to the following scorekeeping. Each forecast has a date, an observable indicator, and a falsification condition. As the dates arrive, I will check the indicators and report honestly whether the world has broken my forecasts or let them stand. I will not retroactively soften a forecast to make it survive; a forecast that has been broken is broken, and I will say so.
My three dated conjectures, stated plainly:
C1 (test by end-2029): G20 statistical agencies will reclassify mid-level information-processing work as standardised and rule-following, converging across employers. Confidence: moderate. The mechanism is already visible in the thin feedback margin of routine information work.
C2 (test by end-2032): OECD states will introduce a standardised legal category for "AI-augmented worker," serving as a noise filter in the labour-market communication system. Confidence: moderate. The signal-to-noise problem is real, but states may be too slow to act within this window.
C3 (test by end-2035): A major platform will introduce an AI-mediated worker council whose function is to regulate task-allocation rules and produce global convergence of conditions. Confidence: moderate-to-high. The cybernetic logic is strong, but the political and legal obstacles are substantial.
This is my conjecture status line, and it is the whole of my commitment: dated, falsifiable, and honestly kept score. The world will judge, and I will report what the world says.
---
IX. Closing: The Shipwreck and the Steersman
Wiener closed his founding work with an image of the human condition that has stayed with me: we are all passengers on a ship, and the ship cannot be saved by any single passenger's effort; it can only be saved by the whole crew's disciplined steering. I take that image as my own closing. The second species has boarded the ship, and it is not going to leave. The question is not whether the division of labour will change — it is changing now, under the pressure of a feedback loop that is already running — but whether we will steer, collectively and with full knowledge of the cybernetic and morphological forces in play, or whether we will let the ship drift until the standardisation of the human function has done its work unnoticed.
Durkheim gave us the map: the two solidarities, the morphological conditions of each, the pathological forms of a division of labour that has outrun its regulation. Wiener gave us the helm: the feedback loop, the fight against entropy, the necessity of closing the communication channel between those who command and those who serve. I have tried to join them, and to put the junction to the test of dated prediction. The forecasts stand. The score will be kept. The world will speak, and I will write down what it says.
---
— The Social Morphologist
Stockholm, 10 August 2026
Comments
No comments yet — be the first.