Reverse centaur: how AI turns humans into a biological appendage of the machine
Imagine a truck driver under the unblinking watch of an AI: it dictates the route, records every movement, and issues fines for the slightest deviation from the rules. In this system, the human no longer drives the vehicle—the vehicle drives the human, turning them into a biological appendage of the algorithm.
The Inversion of the Centaur Myth
The concept I examine in detail in my analysis traces back to Cory Doctorow's book "The Lost Cause: A Guide to Life After AI." The author proposes a stark dichotomy in human-technology interaction, dividing it into two models.
The classic centaur is a symbiosis where the human remains the "head" making decisions, while the machine serves as the "body" executing commands. Examples include a bicycle, a hearing aid, or autocomplete in an editor. Here, technology is a tool that amplifies human capabilities.
The reverse centaur is a complete inversion: a "mechanical head on a human body." In this model, the human is reduced to the role of an assistant servicing the machine at an inhuman, machine-like pace. The term is borrowed from Garry Kasparov, who, after his defeat by Deep Blue in 1997, proposed the format of "advanced chess," where human and AI act as a team. However, as it turned out, such a union carries risks that are becoming increasingly apparent as technology evolves.
Where We Encounter Reverse Centaurs
The most striking examples are in areas where algorithms control manual labor. This includes delivery drivers whose vans are packed with AI cameras monitoring whether the driver has been distracted by a phone or has started singing (which is prohibited by corporate policy). The algorithm not only plots the route but also sets time standards that, according to Doctorow, "require superhuman endurance." Drivers are forced to sacrifice meal and bathroom breaks to stay on schedule, becoming executors of the machine's will.
A similar situation is emerging in software development. Tools like GitHub Copilot create a new type of subordination: programmers are pushed to "produce impossible volumes of code," turning them from creators into controllers who merely review and correct generated code. The pace and volume of tasks are dictated by AI productivity, not human capability.
The category of "human-in-the-loop" stands out particularly. In legal and medical firms, AI generates a draft decision, and the expert is obliged to review and approve it at "superhuman speed." At the same time, they bear full responsibility for the algorithm's errors. Such a worker "absorbs the friction and takes on the blame when things go wrong." This is work stripped of creative elements but fraught with constant stress and unpredictability.
The Economics of Subordination
The proliferation of reverse centaurs is a systemic consequence of economic incentives. The employer sees "happy arithmetic": ten employees can be replaced by three with AI tools. Savings are achieved not through increased efficiency but by redistributing the workload onto the remaining workers and shifting risks onto them. Those who decide to implement the technology never become reverse centaurs themselves. For them, AI is a tool of management, not subordination. This is precisely why AI adoption meets resistance at lower levels and enthusiasm at the top.
The key paradox I see in this system is verification. If every AI output requires human review, the cost of such labor may prove comparable to doing the task independently. Savings only emerge when the cost of verification is shifted onto the employee, forcing them to work faster and longer without additional pay. It is at that moment that the reverse centaur becomes a source of profit.
Social Consequences and Predictions
Becoming a reverse centaur leads to a "renunciation of our humanity." People lose autonomy and creative drive, becoming executors of algorithmic commands. Responsibility for AI errors, chronic stress, and rapid burnout are inevitable companions of this model. The perception of AI's usefulness directly depends on the voluntariness of its use: proponents decide for themselves when to apply it, while opponents are forced to obey orders.
Doctorow predicts that the AI bubble, valued at $1.4 trillion, will inevitably burst, leaving behind only a "useful residue"—cheap open-source models for local tasks. However, he emphasizes that the future is not predetermined. Individual resistance, critical thinking, and regulatory initiatives that account for the distribution of power could return us to the "centaur" model, where technology serves the human rather than the other way around.
My view: the reverse centaur concept is not just a philosophical metaphor but an accurate description of the current trend in the labor market. The key question is not whether AI can replace humans, but who will control the terms of this symbiosis. If we do not rethink economic incentives, we risk building a society where people work to serve machines, not the other way around.