When Work and Power Begin to Separate
How the transformation of labor, ownership, and productivity may reshape the relationship between citizens and economic power and why who controls AI-driven productivity is the question at the heart of “When the West Begins to Resemble Rome”.
There are historical parallels that should be used with caution.
The one between the contemporary West and Imperial Rome is certainly one of them.
To say that “the West is becoming Rome” is an effective phrase, but too generic to be truly useful. Rome was an agricultural society, lacking modern technology, without nation-states, without representative democracy, and without financial markets comparable to those of today.
Its economic, demographic, and institutional world was radically different from ours.
Yet the comparison can become interesting if we stop looking for superficial equivalences and begin to examine the underlying structures.
Not the senator as the equivalent of the billionaire, not the slave as the equivalent of the contemporary worker, not the barbarian as the equivalent of the migrant.
The more serious question is another: what happens when an extremely complex society produces a growing amount of wealth but encounters ever-greater difficulties in transforming it into autonomy, cohesion, security, and political capacity?
This is the starting point of the attached essay, to which I refer you for the complete analysis. And its central argument can be summarized in one sentence: the problem facing the West is not becoming poor, but the risk of progressively separating work, income, property, competence, and power.
For most of the industrial age, however, these elements were closely linked. Work generated income; income enabled consumption and savings; savings could be converted into property; property contributed to economic autonomy; and, finally, work gave citizens social and political clout.
The worker was simultaneously a producer, a consumer, a taxpayer, and a citizen. It was not a perfect system, but it was one in which participation in production constituted one of the main sources of bargaining power.
Globalization, automation, and the digital transformation have begun to separate these elements. A society can consume without directly producing everything it consumes. An investor can share in profits without participating in the production process.
A digital platform can reach millions of people with a relatively small amount of human labor. And artificial intelligence introduces an even more radical change: it does not merely replace some physical labor, but can automate some cognitive tasks.
This is where the issue becomes much bigger than just employment.
According to the International Labor Organization, about one in four workers worldwide is exposed to some degree to generative AI, with a higher proportion in high-income economies.
But the ILO also points out that, as things stand, the transformation of job tasks is more likely than the complete elimination of jobs. The issue, therefore, is not predicting how many jobs will disappear. It is understanding how the productivity gains generated by artificial intelligence will be distributed.
If a technology enables a worker to produce in four hours what previously took eight hours, there are several possibilities. The gain can translate into higher wages, reduced working hours, lower prices, higher profits, dividends, investments, or increased tax revenue. It can also be distributed through a combination of these possibilities.
But something else can also happen: the amount of labor required to produce a given amount of value decreases more rapidly than the system’s ability to redistribute the new surplus.
In that case, the problem would not simply be unemployment. It would be the reduction in the scarcity of human labor.
This distinction is fundamental. A worker does not lose power merely by losing their job. They can also lose it when they become easily replaceable.
If technological capital becomes progressively more productive and human labor progressively less necessary in certain activities, the contractual relationship can change even without an immediate collapse in employment.
And this is where demographics make the issue even more interesting.
Europe is aging. The fertility rate in the European Union fell to 1.34 children per woman in 2024. Italy ranks even lower on the European demographic scale.
This means that in the coming decades, many economies will have to fund and maintain pension systems, healthcare, infrastructure, and public services with a relatively smaller workforce.
Normally, fewer workers mean lower productive capacity. But artificial intelligence could change this equation.
A society with fewer workers can produce more if each worker has access to much more powerful tools. The fundamental equation therefore shifts from simply population × labor to increasingly workers × productivity per worker.
This could be one of the greatest technological opportunities in history. AI could become a way to offset the demographic crisis: fewer people, but greater individual productive capacity.
But there is another possibility.
If the new productivity is concentrated in the hands of those who own the systems, models, data, semiconductors, infrastructure, and capital needed to develop them, the very technology that compensates for demographic trends can increase economic concentration.
This is the central paradox.
Artificial intelligence can be both a mechanism for offsetting demographic trends and a mechanism for concentrating wealth.
The difference will not depend on the technology itself, but on the institutions that determine who reaps its benefits.
This is where the comparison with Rome proves useful.
The essay does not argue that the West is destined to repeat the fall of the Roman Empire. It argues something more cautious: a society can become fragile when its ability to transform surplus into order, security, cohesion, and power diminishes relative to the complexity it must sustain.
The point, then, is not absolute poverty. A contemporary Western citizen can live in material conditions incomparably superior to those of a Roman peasant.
They may own a smartphone, have access to advanced medicines, travel, utilize vast amounts of information, and benefit from services that were unthinkable in ancient times.
And yet they may own an increasingly smaller share of the system that produces that wealth and understand less and less about the technological systems on which they depend.
Here, another dimension of the problem emerges: the cognitive one.
The easier artificial intelligence makes it to obtain answers, the more important it becomes to know how to evaluate them. In a pre-AI society, access to information was often the main limitation.
In an AI-dominated society, the problem may become the opposite: having access to an enormous amount of information without possessing the tools necessary to distinguish what is true from what is plausible but false.
For this reason, the new inequality may not be merely economic. It could concern the ability to understand and control the system itself.
The essay identifies here a possible convergence of three forms of capital: economic, technological, and cognitive.
Those who possess economic capital can finance technology; those who control technology can multiply productivity; those who possess high cognitive skills can understand, govern, and verify it.
If these three forms of capital tend to concentrate in the same hands, the issue becomes far more profound than traditional income inequality.
The real question then becomes: how widely distributed is the capacity to produce, own, understand, and decide?
This, ultimately, is the question that links the discussion of Rome to the issue of artificial intelligence.
The West still possesses enormous compensatory resources: capital, energy, technology, universities, research, financial markets, institutions, tax systems, welfare, military capabilities, and statistical tools with which it can measure its own problems. The difference compared to Rome is enormous.
But recognizing a problem does not automatically mean solving it.
A society may know that its population is aging and still be unable to reverse the trend.
It may know that productivity is growing too slowly and still fail to reform the system.
It may know that debt is limiting fiscal space and fail to reduce it.
It may know that artificial intelligence concentrates power and fail to distribute its benefits sufficiently.
The ability to self-correct is therefore the decisive factor.
And that is why the real question is not whether AI will destroy jobs. It is who will own the productivity that AI will create.
If that productivity leads to more widely distributed capital, more skills, more free time, greater economic security, and broader ownership, artificial intelligence could become the most powerful technological response to the demographic crisis.
If, on the other hand, it increases productivity while further concentrating ownership, capital, and decision-making power, it could produce a society that is materially richer but economically and politically more dependent.
And this is where the comparison with Rome ceases to be a metaphor for decline and becomes a test of the West’s ability to manage its own complexity.
The ultimate question, therefore, is not whether we will become Rome.
It is much simpler and, perhaps, much more difficult: will we be able to use technology to reduce the need for labor without reducing the needs of our citizens?
It is along this dividing line that a decisive part of the West’s next transformation will play out.
Attachment
When the West Begins to Resemble Rome
