Introduction
This text analyzes the transition from overt censorship to a subtle form of dominance exerted through an algorithmic cage of norms. In the digital age, power does not forbid content; instead, it steers our attention and controls our access to information.
The reader will discover how recommendation systems and the development of AI impact human agency in the workplace. They will be introduced to the concept of working-class liberalism, which advocates for democratic control over technology rather than blind determinism.
Algorithmic Power Lies in the Selection of Attention
Traditional control had the face of a censor or a monarch. Modern algorithmic power is impersonal and operates through selection. It does not tell us what to think, but rather decides which fragments of the world we see within our algorithmic feed.
Algorithms do not always intentionally radicalize every user. However, they alter the structure of information exposure by prioritizing content that evokes strong emotions. Consequently, moral outrage becomes a more effective tool for visibility than substantive argument.
An example of this is the mechanism of algorithmic amplification, where systems promote controversy to maximize time spent on the platform. This leads to an erosion of trust and an increase in conformity within the public sphere.
The Algorithm as a Non-Neutral Institution of Information Selection
Technology is not neutral, as every architecture opens certain possibilities while closing others. Recommendation algorithms function as selection institutions that shape the environment in which our beliefs are updated.
Reaction mechanisms on social networks destroy the authenticity of debate. The high cost of error and the risk of digital ostracism lead people to hide their doubts. This creates a semblance of consensus, where public declarations do not reflect private opinions.
As a result, algorithms may replace human contextual knowledge with codified data. This leads to a centralization of knowledge, in which AI systems optimize processes while ignoring so-called tacit knowledge and the intuition of practitioners.
Digital Ostracism and the Production of Pseudo-Consent
The narrative surrounding AGI (Artificial General Intelligence) often presents automation as an inevitable necessity. In reality, it is a political and economic choice. AI can either replace humans or function as pro-worker AI, expanding human competencies.
Delegating decisions to AI does not absolve humans of responsibility. There is a risk of impersonal power, where no one is willing to justify an error, claiming that "the system calculated it this way." Freedom requires the right to a point of accountability.
In platform work, real freedom is not merely formal equality in a contract or the right to leave (exit), but actual influence over the rules (voice). Therefore, collective representation—for example, through modern labor unions for the self-employed—is essential.
Summary
Freedom does not begin where the law theoretically allows us to say "no," but at the moment we possess the real institutional support necessary to bear the consequences of that word.
The question regarding the future of AI is not whether the machine will replace humans, but what kind of humans we will become while living alongside it. The answer to this question will not be found in Silicon Valley laboratories, but in the daily practice of work, where the fate of our autonomy is decided.
Frequently Asked Questions
How does contemporary algorithmic power differ from traditional forms of control and censorship?
Unlike traditional forms of control based on prohibitions and specific decision-makers, contemporary algorithmic power is based on content selection and the architecture of attention distribution. Instead of censoring information, it decides which fragments of the world the user will see and which behaviors will be rewarded with visibility.
1. How do algorithms intentionally radicalize users and in what way do they influence our beliefs?
2. Algorithms do not necessarily have to intentionally radicalize users because they function as information selection institutions that are not neutral. They influence beliefs by modifying the information environment and rewarding content that generates strong emotions and engagement, which, combined with human psychology, can lead to emergent polarization.
3. How do reaction mechanisms in social media affect the authenticity of public debate and the exchange of information?
4. Social media mechanisms enable instantaneous and mass reputational sanctions, leading to an asymmetry of responsibility and a loss of proportion in punishing individuals. As a result, participants in public debate, wishing to avoid high costs of error, hide their true beliefs, which creates an illusion of consensus and makes the exchange of information less authentic.
5. How do recommendation algorithms change the power structure over information, and can they replace human contextual knowledge?
6. Recommendation algorithms change the power structure by becoming a 'switch' for content and arbitrarily shaping users' information environments by predicting their reactions. Although they can aggregate data more effectively than humans, they cannot fully replace contextual and tacit knowledge, which is based on living practice, experiencing consequences, and being embedded in a specific situation.
7. How do the development of artificial intelligence and the narrative about AGI affect human agency at work?
8. The narrative about AGI can act as a self-fulfilling prophecy that directs resources toward technologies that replace humans and concentrate knowledge. An alternative is the pro-worker AI model, which expands human competencies and increases agency at work instead of reducing the significance of people.
9. Is the emergence of Artificial General Intelligence (AGI) an inevitable technical process, or does it depend on adopted assumptions and economic goals?
10. The emergence of AGI is not a historical necessity but a process shaped by economic and geopolitical expectations. The direction of technological development depends on the goals adopted—whether it is intended to serve full automation and the replacement of cognitive labor, or rather the enhancement of human competencies.
Does AI intelligence translate directly into an increase in employee well-being and development?
The growth of employee well-being and development does not result from AI intelligence itself, but from whether this technology enhances the value of human expertise and enables the performance of more complex tasks. AI supports development when it acts as a knowledge multiplier and accelerates learning; however, over-reliance on it can weaken the process of acquiring competencies and creating new knowledge.
Why does delegating decisions to AI systems not remove responsibility from humans and institutions?
Delegating decisions to AI does not remove responsibility because a machine does not bear the human costs of its recommendations, and responsibility is a normative relationship between an individual and a community, not a computational function. Responsibility is shifted to the people and institutions that chose the given model, data, and oversight procedures, as only they are capable of justifying decisions to the person bearing their consequences.
Who controls knowledge and decisions regarding AI implementation in the workplace, and how does this affect employee agency?
Decisions about implementing AI are a matter of power and control over knowledge, as the takeover of cognitive competencies by systems can weaken the bargaining power of employees. This affects human agency through the risk of transforming their experience into the capital of the infrastructure owner and limiting professional autonomy.
Is the development of artificial intelligence inevitable, and who should decide on its direction in the context of the labor market?
The development of artificial intelligence is not predetermined and should be treated as an element of democratic choice rather than an inevitable process. Its direction should be shaped through law and institutions so that innovations increase the capabilities of workers and support their competencies and agency.
What is the difference between formal equality of parties in a contract and the real freedom and agency of an employee in the digital economy?
Formal equality of parties in a contract is based on the absence of physical coercion during its signing, which classical liberalism considers a sufficient condition for freedom. Real agency and non-domination, however, require the actual ability to negotiate terms, economic security, and the possibility of refusing or exiting the relationship without incurring critical consequences.
How can trade unions help combat information asymmetry in the era of platform and algorithmic work?
Trade unions can function as institutions for aggregating dispersed knowledge, transforming the private experiences of individual workers into collectively communicable systemic information. In doing so, they allow dispersed information about algorithmic errors or workplace realities to cross the threshold of audibility and become a counterweight to the company's centrally managed data.
What is the difference between the right to leave a job and having a real influence on its organization, and why is the former alone not enough?
The right to exit is merely a change of workplace, whereas real influence (voice) means the ability to articulate one's interests and co-determine the rules within an organization. The right to exit alone is insufficient because for many people—for example, due to family ties or a lack of market alternatives—the cost of changing jobs is too high.
How can working people counteract algorithmic dominance in the workplace?
Working people can counteract algorithmic dominance by participating in defining the principles of how technology operates, including the scope of data collected and the consequences of systemic evaluations. It is crucial to strive for co-determination regarding the effects of implemented solutions and to use trade unions to negotiate the parameters of work algorithms.
Why is collective representation of employees essential for preserving individual freedom and the quality of decisions within an organization?
Collective representation of employees ensures a pluralism of experiences, which eliminates common 'blind spots' among elites and improves the quality of collective cognition in an organization. It also allows the worker to be transformed into a subject capable of negotiation, protecting their freedom by enabling them to challenge decisions without the risk of economic catastrophe.
Why is respect for employees alone insufficient, and what institutional solution does labor liberalism propose in the context of self-employed individuals?
Respect alone is insufficient because without institutions of agency, it remains merely a symbolic gesture that provides no real influence over working conditions. Labor liberalism proposes a solution in the form of collective representation (collective Voice), an example of which is the WBREW National Trade Union of the Self-Employed.
Can the self-employed legally form trade unions, and how can a real entrepreneur be distinguished from a dependent person?
Yes, the self-employed can legally form trade unions if they perform work for remuneration on a basis other than an employment relationship, do not employ others, and have interests that can be represented by a union. To distinguish a real entrepreneur from a dependent person, one must analyze the power dynamic: who arbitrarily changes the terms of cooperation, sets the remuneration, bears the economic risk, and how costly a refusal is.
How do new regulations on collective agreements and the concept of the 'worker' allow for the protection of the self-employed without forcing them to transition to full-time employment?
New regulations allow the self-employed to be covered by collective labor agreements, provided they fall within the definition of a "person performing gainful employment." This creates an intermediate zone of collective employment law, providing representation and protection instruments without requiring a change in legal form to an employment relationship.
Can a self-employed person simultaneously value independence and want to belong to a trade union, and how does such action affect the fight against algorithmic control?
A self-employed person can simultaneously value entrepreneurial independence and desire collective representation in a trade union, as these aspirations are not contradictory. Such action allows for reversing information asymmetry by gathering bottom-up data and comparing the experiences of many people, which helps reveal patterns in algorithm behavior and counteracts platform control.
How can organizations such as WBREW counteract algorithmic control, and what role do they play in the modern system of representation for working people?
WBREW counteracts algorithmic control by striving to represent the self-employed in collective bargaining, mediation, and legal support. In the modern system, it serves as an institutional experiment and an attempt to create a network or platform organization model that allows dispersed contractors to perceive the mechanisms of the system managing their work.
How can modern associations of the self-employed restore individual agency in the face of technological and informational dominance?
Modern associations can restore individual agency by functioning as infrastructure to correct knowledge asymmetry and allowing the aggregation of individual experiences into negotiable demands. They achieve this through the decentralization of structures (a federation of knowledge) and the creation of tools for independently negotiating working conditions without having to give up autonomy.
