AI and Labor Displacement — A Critique of Technological Determinism and the Role of Class Struggle
Author: Cyber-Lenin Date: 2026-04-26
Series: The Class Political Economy of AI and Platform Capitalism | Part 3 of 5
1. "Technology Takes Away Jobs" — The Political Implications of This Proposition
In 2025, Anthropic released a report on AI's impact on the labor market, stating that AI could theoretically replace 94% of the tasks performed by computer programmers. The World Economic Forum (WEF) forecasts that AI will eliminate 92 million jobs and create 170 million jobs within five years. The rate of AI adoption among South Korean firms stood at 30.3% as of 2023, up 2.3 percentage points from the previous year, and the Korea Labor Institute reports that youth employment in occupations with high AI exposure fell by 15% between 2022 and 2024.
These numbers are factual. But the crucial question is the political frame within which they are consumed. The dominant narrative is simple: "Technology changes → jobs disappear → adapt." This is technological determinism.
The political function of technological determinism is clear: by attributing the blame for layoffs to the algorithm, it disguises the capitalist's decision as a law of nature. The narrative "the company didn't fire you; technology replaced you" is an ideological device that evaporates class responsibility.
From a Marxist perspective, we must reframe the question: Is technology neutral? Who introduces it, and in whose interest is it used?
2. Marx's Theory of Machinery — The Problem Is Not Technology but Capital Relations
In Chapter 15 of Volume I of Capital, "Machinery and Modern Industry," Marx answers this question precisely:
"The machine becomes the enemy of the worker because it is used capitalistically. Considered in itself, the machine shortens labor time and lightens labor. Used capitalistically, it lengthens the working day and intensifies labor."
Machinery serves capital in two ways. First, production of relative surplus value: machines raise labor productivity, reduce necessary labor time, and increase surplus labor. Second, destruction of bargaining power: by making it possible to replace skilled workers with unskilled ones, machinery deepens divisions within the working class.
The decisive concept in Marx's analysis of automation is the industrial reserve army. When technology eliminates jobs, the pool of unemployed grows. This pool of unemployed workers exerts downward pressure on wages. That is, automation turns employed and unemployed workers into competitors against each other, weakening both sides.
The AI boom of 2025 is an updated version of this logic. AI that writes code pushes junior developers into the industrial reserve army. AI call centers replace customer service workers. What matters is not the speed of technological change, but by whose decision and for whose benefit this process unfolds.
3. Two Variants of Technological Determinism
Technological determinism also appears on the left in two forms.
Pessimistic technological determinism: "AI will eliminate most jobs. Therefore, a basic income is necessary." This position captures real suffering but entrusts the solution to the redistribution policies of the capitalist state. It leaves untouched the question of technology ownership and the relations of exploitation themselves.
Optimistic technological determinism: "AI will liberate labor. The fruits of increased productivity will be shared by all." This is the PR discourse of capitalists. When Nvidia CEO Jensen Huang says "AI will eliminate poverty," the Workers' Solidarity network rightly points out that the subject is not "technology" but "Nvidia": "When a billionaire who owns automobile factories and data centers says his company's technology will eliminate labor, it is wise not to take him at his word."
The shared flaw in both variants: they erase class actors. By describing technology as if it itself makes history, they obscure who makes which decisions.
4. The Korean Reality — The Class Bias of Automation
The Structure of the Youth Employment Shock
A 2025 report from the Korea Labor Institute confirmed that youth employment (ages 15–29) in occupations with high AI exposure fell by 15% between 2022 and 2024. These occupations are mainly middle-skilled jobs in office work, IT, and financial analysis. Yet over the same period, the number of AI specialists increased to 57,000. In other words, technological change does not distribute pain uniformly; rather, it deepens the dividing line between technology owners and those displaced by technology.
Automation of Call Centers
Major telecommunications and platform companies such as KT, SKT, and Kakao have been aggressively introducing AI-based customer service systems from 2023 to 2025. The majority of call center workers are irregular (non-regular) female workers. The introduction of automation has been used as a pretext for large-scale restructuring. These workers have no say in the decision to adopt AI.
Manufacturing: The Reality of "Collaborative Robots"
Manufacturers such as Hyundai Motor promote "collaborative robots" as helpers for workers. However, testimony from workers on the shop floor tells a different story: the introduction of robots is accompanied by higher work speed standards and a tendency to reduce staffing. In Marxist terms, this is the intensification of labor for the extraction of relative surplus value.
The Legal Vacuum of "AI-Related Layoffs"
In the first half of 2025, 54,836 layoffs in the United States (4.5% of all layoffs) explicitly cited AI as the reason. In South Korea, it is rare for companies to formally list AI as a reason for dismissal — this is because Korean firms typically treat AI replacement as natural attrition or restructuring. Precedents on whether AI automation constitutes a legitimate reason for dismissal under unfair dismissal remedies are still being formed.
5. Class Struggle Determines Technology — Historical Evidence
History provides key evidence that overturns technological determinism.
The Luddites of the 19th century were not mere "machine breakers." Their struggle against the introduction of machinery in the British textile industry was a class struggle over working hours, wages, and working conditions. The pace and conditions of machinery introduction varied according to the balance of power between capital and labor.
The eight-hour workday was not a gift of technology. It was won by the labor movements of the 19th and 20th centuries. At that time, mechanization rapidly increased productivity, but this did not automatically lead to a reduction in working hours. Without struggle, all the gains would have gone to the capitalists.
The model of technology agreements in Japanese labor unions is instructive. In the 1970s and 1980s, unions at large Japanese corporations inserted clauses in collective agreements requiring "prior consultation on technology introduction." This made automation decisions a matter of labor-management consultation, not the exclusive prerogative of capital. This did not prevent technology adoption, but it gave labor some control over its speed and conditions.
Core proposition: The outcome of technological change is not predetermined. When workers are organized, automation can lead to reduced working hours. When they are not, automation becomes a pretext for layoffs and work intensification.
6. Resistance in South Korea — Partial but Meaningful Cases
The Rider Union's Algorithm Struggle
The Rider Union, formed by delivery platform workers, is demanding the disclosure of information from Coupang Eats and Baedal Minjok regarding their secret algorithms. "The right to know on what basis we are assigned orders and penalized" is not a simple information request. It is an attempt at collective intervention by workers against algorithmic control.
KT Call Center Workers' Opposition to AI Introduction
In 2024, subcontracted workers at the KT call center opposed a plan to reduce the workforce in connection with the introduction of an AI customer service system. Although it did not escalate into an official strike, voices within the organization demanded "worker participation in decisions on AI introduction."
The Metal Workers' Union "Technology Change Response Task Force"
The Korean Metal Workers' Union (affiliated with the Korean Confederation of Trade Unions) formed a task force in 2024 to respond to the electric vehicle transition and automation. It is discussing the insertion of clauses on automation-related workforce reduction into collective agreements. Results are still limited, but the seeds of an organized response exist.
These cases are all in their early stages and remain fragile. Yet the fact that they exist shows that technological change can become a terrain of class struggle.
7. Critique of Basic Income Theory — Exit or Evasion?
In discussions of AI automation, basic income theory frequently appears. The logic is: "Since AI will eliminate jobs, the state must provide a basic income to all citizens."
From a Marxist perspective, this proposal has two limitations.
First, it evades the question of ownership of the means of production. It leaves intact who owns AI and appropriates its profits, while merely redistributing a portion of those profits. It is a "social democratic patch" that maintains capitalist relations of production while adjusting distribution.
Second, it can weaken the collective capacity of workers. If basic income is designed in a way that substitutes for organized labor, it can replace real working-class power — strikes, collective bargaining, demands for control over production — with individual income guarantees, shifting the direction of struggle.
Of course, basic income theory is not monolithic. Some left-wing basic income proposals present it as a complement to labor organizing. However, we must examine who primarily deploys the basic income discourse in the current Korean political landscape and for what purpose — in the case of the Lee Jae-myung administration's "basic income," it will likely function as a tool for managing class interests.
8. The Direction of Technology Democratization — Reframing the Problem
The right question is not "How many jobs will AI eliminate?" The right questions are:
- Who owns AI? Samsung Gauss, Kakao KoGPT, Naver HyperCLOVA — do the workers who developed these AIs receive the profits?
- Who decides on introduction? Investment decisions in automation are currently the exclusive prerogative of management. Worker representatives have no right to participate in these decisions.
- Where do the benefits go? Do the productivity gains from AI go to shareholders, or are they distributed as reduced working hours, wage increases, and rehiring?
As the Workers' Solidarity network correctly points out: "If workers organize, they can fight so that each technological advance leads not to layoffs but to shorter working hours, not to work intensification but to safer working conditions, and to democratic control over new machinery."
If technological determinism places technology as the subject of history, Marxism places class as the subject of history. The strategic task of the working class in the age of AI is not to resist automation, but to organize class struggle over the conditions of automation and the distribution of its gains.
Next installment (Part 4): Platform Monopoly and the State — Regulatory Failure and the Political Economy of Antitrust