The emergence of Artificial Intelligence (AI) and automation is profoundly transforming the economy, employment, and the functioning of the industrial production system. The question that is increasingly gaining traction is simple yet unsettling. If machines perform more and more human work, who will pay the taxes that sustain the Welfare State?
This debate on taxation and employment comes amid the rapid expansion of the AI era. The current tax system mainly relies on three sources: taxes on labor, consumption, and capital. But in many countries, revenue from labor — social security contributions and income taxes — remains a fundamental part of public financing. Moreover, an aggravating factor is that this technology is primarily in private hands.
It is obvious that digital technologies have the potential to increase productivity and generate wealth, but they can also concentrate economic power and profoundly transform the labor market. Therefore, many experts agree that the political decisions made in the coming years will be crucial in determining how the benefits of the technological revolution are distributed. Hence, many experts warn that if human employment decreases as a result of automation, the state’s funding model could be jeopardized.
The recent and highly commendable encyclical Magnifica Humanitas by Pope Leo XIV on the safeguarding of the human person in the age of artificial intelligence warns: “Throughout the centuries, technological development has contributed to a significant improvement in humanity’s living conditions; at the same time, each stage of progress has also revealed the ambiguous side of tools capable of causing harm when not oriented toward the good.”
“Today, however, we face a new situation, in which the power and ubiquity of emerging technologies intertwine with the fabric of daily life, shape decision-making processes, and profoundly impact the collective imagination: ‘Never has humanity had so much power over itself.’ New technologies open a horizon that extends in directions that, though intuitive, we cannot yet fully foresee. This makes it more complex to assess their impact and long-term effects on human dignity and the common good,” he adds.
This accelerated deployment of generative AI tools in the workplace has also reopened a classic debate from a Marxist perspective: Who truly benefits from the productivity increase brought about by technology? The worker, who may see their workload reduced and their creative capacity enhanced, or the company, which capitalizes on this efficiency gain in the form of surplus value? This question is not new; it is part of a long tradition of Marxist analysis of the relationship between technology, labor, and capital.
From a Marxist view, technology is part of the productive forces — that is, the material and technical elements that determine how social wealth is produced. Generative AI greatly expands these forces: it automates cognitive tasks, accelerates processes that were previously intensive in time and knowledge, and allows a worker to achieve higher-quality results in less time.
However, for Marx, productive forces never act alone; they are embedded in social relations of production, and it is this structure that decides how their benefits are distributed. In a capitalist system, technological improvement tends to be appropriated by capital — not by labor — because increased productivity does not automatically translate into better wages, shorter working hours, or greater autonomy for the worker. In the case of AI, as with many historical examples, it also becomes a threat of substitution.
From a Marxist conception, technological progress has historically been used to replace labor, creating a reserve army of the unemployed that puts downward pressure on wages. The labor taxes previously paid by displaced workers — or reduced due to precarious employment — must be compensated.
Faced with this, Marxism proposes that tax collection should fall directly on the massive profits of big tech capital, rather than taxing consumption or the income of working families. Through a Marxist lens, artificial intelligence does not pay taxes; the exploited working class pays them. From this perspective, AI is “constant capital” (machinery and technology).
Therefore, labor taxes and the extraordinary profits derived from automation are extracted from the surplus value generated by workers. But in reality, Artificial Intelligence (AI) itself does not pay labor taxes, as it lacks legal personality. It is companies and employers who pay taxes and social security contributions. Given that AI replaces or automates human jobs, the current debate proposes that companies pay specific taxes to compensate for the loss of tax revenue from displaced labor.
The proposal to tax automation draws on previous international debates. In the United States, for example, Senator Bernie Sanders once advocated for similar measures to compensate for job destruction caused by robotization. Public figures like Bill Gates have supported similar ideas.
The central argument is that technological advances increase productivity and generate huge corporate profits, but that money does not always translate into better wages or more jobs. Therefore, a tax system adapted to the digital economy could help redistribute part of that value.
In the current economic debate, figures aligned with protectionist approaches have proposed so-called “robot taxes” or AI taxes. The idea is that if an algorithm replaces a human, that machine or the owning company should pay an amount equivalent to what the worker previously contributed.
The goal is not to stifle innovation, but rather: to redistribute automated wealth. To prevent tech corporations from monopolizing the productivity gains, therefore, any tax on AI is actually a tax on the capital accumulated by entrepreneurs. However, we greatly fear that the dilemma is not technological but social: Will AI serve to free up life time or to intensify exploitation? The answer will depend on who controls the algorithmic means, but if there is one conclusion we must draw, it is that AI is not neutral — that is the reality of this dystopian world.
Uruguayan journalist resident in Geneva, former member of the United Nations Press Correspondents Association in Geneva, analyst Associated with the Latin American Center for Strategic Analysis (CLAE)