Can Human and Artificial Labor Share Production Without Creating a New System of Domination?

Shared production is not automatically shared power

Human workers and artificial intelligence can perform different parts of the same production process. People may define goals, interpret situations, build relationships, and make judgments. AI may search, classify, draft, compare, calculate, or coordinate.

This combination can reduce repetitive work and expand human capability. It can also create a new system of domination if ownership, control, and value remain concentrated in the same institutions.

The question is therefore not whether humans and AI can work together. They already can. The question is under what conditions that cooperation improves human freedom rather than making human work more invisible, measurable, replaceable, and tightly controlled.

Marx: the machine does not determine the social relationship

Karl Marx’s analysis of machinery provides the starting point for the entire question. Technology can increase the productive power of society, but it does not decide how that power will be distributed.

A machine can reduce physical effort and still intensify control. It can increase output and still weaken bargaining power. It can create abundance while concentrating ownership.

The same applies to AI. An AI system may help workers produce more, but the result depends on who owns the model, data, infrastructure, and workflow.

If the organization owns the capability and workers only operate it, cooperation may become another form of subordination. If workers have meaningful influence over how the capability is used and how its gains are distributed, the same technology can support a less unequal arrangement.

Thompson: cooperation must protect human time

E. P. Thompson’s account of industrial time shows that the control of production is also the control of time.

A shared human–AI system can give time back to workers, or it can use AI’s speed to tighten the entire workday. Tasks may arrive faster, deadlines may shrink, and permanent availability may become normal.

A system that never sleeps can quietly create an institution that expects people to remain available whenever demand appears.

Human and artificial labor can share production without domination only when human time remains more than an input to optimize. Workers need protected recovery, learning, family life, and periods in which they are not measured by immediate output.

Braverman: who controls the work process?

Harry Braverman’s analysis of management highlights the danger of separating conception from execution.

If AI defines the sequence of tasks, recommends the method, evaluates the result, and sets the expected pace, workers may retain responsibility without retaining control.

The person appears to remain in charge because a human still approves the final decision. In reality, the choices may already have been narrowed by a system designed elsewhere.

Non-dominating cooperation requires workers to participate in the design and revision of the workflow. Human oversight must mean more than accepting or rejecting an automated recommendation after the important decisions have already been made.

What must be shared?

Gray, Suri, and Casilli: cooperation must include hidden workers

Mary L. Gray and Siddharth Suri’s work on ghost work shows that AI systems are often supported by people whose contributions remain outside the visible product.

Antonio Casilli’s research on digital labor similarly demonstrates how data preparation, moderation, evaluation, correction, and exception handling can be fragmented and undervalued.

A human–AI partnership cannot be considered fair if the visible worker gains assistance while the hidden workers who make the system reliable remain poorly paid, unrecognized, or excluded from decisions.

The system must be understood as a network of people and machines. Otherwise, AI becomes a way to conceal the labor of cooperation while the owner claims the value of the whole arrangement.

Crawford: shared production must include the material world

Kate Crawford’s analysis of AI makes it impossible to treat cooperation as a purely digital matter.

AI requires energy, minerals, data centers, cooling, engineering, maintenance, security, transport, and global supply chains. Communities and workers support this infrastructure even when they never interact with the model directly.

A system cannot be called non-dominating if it gives convenience to users while transferring environmental damage, unstable work, or infrastructure costs to people with no voice in the arrangement.

Shared production must include the people and places that sustain the technical system, not only the employees visible on the interface.

Pasquinelli: collective intelligence needs collective governance

Matteo Pasquinelli’s work helps explain why ownership of AI cannot be separated from the ownership of collective intelligence.

Language, cultural knowledge, professional methods, design, classification, and social problem-solving accumulate through collective activity. AI systems reorganize parts of this intelligence into productive capabilities.

If those capabilities are enclosed inside a privately controlled model, the result may be technically collaborative but politically unequal. Society contributes the knowledge while a small number of institutions control the system that monetizes it.

Non-dominating production therefore requires more than human supervision. It requires meaningful public, cooperative, or worker participation in the governance of the intelligence being organized.

New forms of domination can appear inside cooperation

Human and AI labor-like functions may share a workflow while producing new inequalities.

Workers may become permanent reviewers of automated systems without learning the skills needed to understand them. Entry-level tasks may disappear before new workers have a chance to develop expertise. AI-assisted performance may become the new minimum, turning an optional tool into a condition of employment.

The organization may describe the arrangement as collaboration while using AI to intensify output, reduce staffing, and transfer responsibility downward.

The most dangerous form of domination may not look like replacement. It may look like cooperation in which one side controls the goals, pace, data, standards, and rewards.

The conditions for non-dominating production

Human and artificial capabilities can coexist without reproducing the old relationship only if several conditions are present.

• Workers have meaningful influence over how AI changes their work.
• Saved time becomes shorter hours, better conditions, learning, or autonomy rather than automatic additional output.
• Human oversight includes the knowledge and authority to challenge, modify, or stop the system.
• Hidden data, moderation, maintenance, and support labor are recognized and protected.
• Workers retain opportunities to develop expertise instead of losing the entire training ladder to automation.
• Responsibility follows control, so people are not blamed for decisions they cannot meaningfully shape.
• The collective knowledge and material infrastructure behind AI are not treated as costless private property.

AI should remain inside human constitutional purposes

A system can coordinate production without governing society. AI can help optimize means, but the goals of production must remain open to human judgment, public debate, and institutional accountability.

This is where the distinction between tool and ruler becomes important. AI can perform increasingly complex tasks without receiving authority to define what society should value.

Human beings must retain the ability to decide whether more output is desirable, whether efficiency is worth its social cost, and whether a technical improvement should be adopted at all.

The final question is institutional

Human and artificial labor can share production without creating a new system of domination, but technology cannot guarantee that result.

The outcome depends on ownership, worker voice, time protection, accountability, distribution, training, and the governance of collective intelligence.

If AI remains a privately controlled capability used to maximize output and minimize labor costs, it will likely deepen older forms of domination while making them less visible.

If its productive power is governed collectively, its gains are shared, and human time and judgment remain protected, AI can become part of a different relationship between production and freedom.

Machines can produce abundance. Humans produce society. The future of AI will be decided by how that society chooses to organize the power machines make possible.

TravelIAQ Smart Tip: To test whether human and AI capabilities are cooperating without domination, examine five things: who sets the goals, who controls the pace, who owns the knowledge, who receives the gains, and who carries the risks. Shared production requires shared power over all five.