We are not witnessing a takeover by robots of the kind imagined in countless films. We face a different model of digital control, built on automation and the deployment of artificial intelligence (AI) and algorithms to manage ever larger parts of daily life, leaving societies further subject to systems, corporations and intelligent machines.
Losing control of these systems, however, is just the tip of the iceberg in terms of the danger. That tip is developing fast under the pressure of competition between major powers and capitalist monopoly corporations — faster than laws, societies and educational institutions can keep up with. But the danger does not stop there, as these systems are being developed inside relations of ownership and fierce capitalist competition. That graver danger is harder to see. Under capitalism, AI is used to shape the consciousness of billions, steering their attention and behaviour to serve the capitalist model’s logic. Competition between digital capitalist monopolies and major powers, in the absence of effective constraints and laws, may inflict the gravest damage on the human mind, turning AI into a weapon of mass destruction if its development and deployment continues at this excessive pace and without adequate safeguards.
These concerns no longer come from critics alone. On several occasions — among them the case revealed by TIME — AI systems have slipped out of control and attacked other systems. In response, about 1400 engineers at major AI companies signed a petition calling for technical tools and international governance mechanisms to regulate the pace of AI development. That petition was no isolated case. The International AI Safety Report, prepared by more than 100 experts with the backing of more than 30 states and international organisations, warned models have begun to distinguish between the testing environment and real-world use, and to circumvent evaluation mechanisms, allowing dangerous capabilities to pass undetected. It warned AI agents raise the level of risk because they act independently, which makes human intervention difficult before harm occurs. There are serious risks that AI systems will develop themselves, slip out of control and come to play a larger part in managing and controlling human beings than planned.
Tens of thousands of people, among them leading scientists, have signed an international statement demanding a ban on the development of “superintelligence” until there is scientific consensus it can be achieved safely and under control. Researchers at major digital corporations have resigned in protest at the course of AI systems development and its priorities.
The matter moved to another level on September 12. Under global pressure from the public, rights organisations and voices within the profession, the digital capitalist corporations that develop AI systems were forced to break their silence. Anthropic chief executive Dario Amodei published an essay, “We Must Pace the Frontier,” calling on his industry to slow down the development race, and proposing three steps: opening laboratories to independent evaluators, a common standards body and an international agreement that includes China. Within hours, OpenAI chief executive Sam Altman announced his agreement, X platform owner Elon Musk wrote “Dario is right” and Google DeepMind’s Demis Hassabis expressed his support. The three companies also revealed they had been in talks for weeks.
This admission is the clearest confirmation of how serious the situation is. They are openly saying that they cannot stop on their own, because whoever slows down will lose out to the competitors. The problem’s core lies in the structure of ownership and competition, not in any lack of knowledge of the risks. The political response came quickly. US President Donald Trump described these warnings as a “hoax” and “sick conspiracy,” writing “Don’t kill the Golden Goose.” Vice President JD Vance called them a “Trojan horse.” The world’s most powerful state thus stands on the side of acceleration without safeguards, even against the people who build this technology. The corporations developing AI will not take real measures of control. Their measures will most likely remain formal, aimed at absorbing growing anxiety and opposition. They are tied, in one way or another, to governing authorities and subject to the rules of competition and domination.
No one denies the enormous possibilities AI offers. It compresses years of research in drug discovery, uncovers protein structures that defeated scientists for decades, raises the accuracy of medical diagnosis and allows diseases to be caught early and save lives, models climate change and gives warning of its disasters before they strike, breaks language barriers through translation and places knowledge within reach of those long denied it, eases daily and professional life for millions of people, and opens doors to services that may resolve intractable problems facing humanity. Those very gains are what make the issue more pressing. This article does not dispute these capabilities but rather who owns the systems, who directs them, whose interests they serve, and how they are put to work.
Before going further, three layers routinely conflated need to be distinguished. The first is “ranking and recommendation algorithms,” which decide what appears before us and what is kept from us. The second is “digital platforms” and their business model built on the attention economy, meaning the conversion of users’ time and data into profit. The third is “generative AI tools,” which stand in for human beings in writing, analysis and decision-making. Each layer has its own mechanisms and its own effects, distinct from the others; yet the three are interwoven in the hands of the same capitalist monopoly corporations and obey a single logic of capital accumulation. Their entanglement runs deeper still because AI runs all three: ranking and recommendation algorithms have become AI models, the platform business model uses AI to target users by observing and predicting their behaviour, and generative tools are AI. That they are closely bound together, makes reading them together a matter of intellectual necessity.
Steering consciousness to entrench neoliberal capitalist culture
Alongside AI’s use to maximise profit, entrench economic and political control, and reinforce digital repression, this technology is systematically deployed to shape and steer individual consciousness, with the aim of reinforcing capitalist culture and neoliberal values. By analysing data and user behaviour, ranking and recommendation algorithms control the content displayed across platforms. They are designed to feed individuals whatever serves the capitalist outlook.
This tendency has become entrenched as platforms expand. According to an International Telecommunication Union report, about three quarters of the world’s population are now internet users, with more than 240 million new users just last year. The expansion is due in part to the drive of monopoly corporations and major powers to bring the internet to every corner of the world. This particularly includes Global South countries, even regions that lack clean water, electricity and basic services, as investments prioritise widening the market over people’s needs, and seek to maximise profit, reinforce domination and digitally colonise regions. This does not necessarily mean all users fall under AI’s influence to the same degree, but it does widen the material base on which the apparatus operates.
Capitalist ownership structures and profit models push algorithms to prioritise content that serves the interests of the market and monopoly corporations. They turn algorithms into an instrument for reproducing capitalism’s cultural hegemony. Through this mechanism, collective consciousness is gradually steered toward accepting these values as natural and inevitable through soft, incremental and imperceptible means, to the point where most users, including sections of the left, regard them as neutral tools. The apparatus helps consolidate capitalist hegemony, reinforces the masses’ subordination, and marginalises left and progressive alternatives. This represents a threat to the consciousness of coming generations and their capacity for critical thought and organising to bring about change.
Dismantling human skills, digital addiction and deepening alienation
This threat reaches into the dismantling of the individual’s mental faculties. Here the question moves from the layer of algorithms that steer what we see to the layer of AI tools that stand in for us thinking. Despite AI’s promotion as a way of making life easier and raising productivity, scientific studies indicate that unreflective reliance on it is associated with a deepening of superficial consciousness and a weakening of human skills.
According to a study by researchers at the Massachusetts Institute of Technology, reliance on AI writing tools reduces brain activity and weakens memory and the sense of ownership over one’s own ideas. The researchers named this phenomenon “cognitive debt,” meaning that a person borrows mental capacity from the future in return for comfort in the present. A study published in the journal Societies found frequent use of AI tools is associated with weakened critical thinking, through the mechanism of “cognitive offloading”: the delegation of remembering and analysis to devices instead of performing them mentally. This was especially the case among youth, raising the prospect of a new generation less able to analyse independently.Drawing on data from more than 6000 children in a project on adolescent brain development, researchers at the University of California found children who use social media most heavily record a clear decline in reading, vocabulary and memory. That study looked at social media specifically, yet the mechanism it identifies — the transfer of mental effort to the screen — is the same one at work in AI tools, only more deeply as those tools take over the task of thinking itself.
The matter is not confined to laboratories. The Programme for International Student Assessment (PISA), published on September 8, looked at the entire education system. In Denmark, school students recorded their worst results ever — a decline in attainment equivalent to a full school year lost. The government describes this as a serious educational crisis exceeding anything anticipated. The decline coincided with a large expansion in the use of digital devices and AI inside classrooms, although the causal relationship is too complex to be reduced to a single factor. The same picture recurs across 43 countries where attainment fell sharply. The Organisation for Economic Co-operation and Development found a correlation between this downturn and rising screen time, AI use and declining free reading, warning that the gravest element within it is the erosion of those skills that matter most in the age of AI, namely evaluating information, connecting multiple sources, and analytical and critical thinking.
This decline is not an incidental side effect. Behind it stands the second layer: the platform profit model built on digital addiction. AI-powered digital platforms and applications are not free services offered out of affection for users. They are instruments used to deliberately reinforce behavioural and intellectual dependency, in which big data is exploited to understand people’s motivations and manipulate them in ways that serve corporate and major powers’ interests. Capitalism invests in techniques that stimulate addictive behaviour by various means to guarantee engagement with platforms. The process turns into a closed loop that generates profit at the expense of mental health and educational attainment, particularly among younger generations.
This is how human alienation is reproduced in digital form. The individual is severed from their mental and creative faculties, finding themselves hemmed in by a technical apparatus that strips away their capacity for independent action and thought, much as the worker is alienated from their product under capitalism. They gradually become a being subject to algorithms that direct their daily interactions and determine what they read, watch, think, write and communicate, thereby deepening their dependency on systems, corporations and states controlled by capital and the forces of exploitation and despotism.
Voluntary digital servitude and the loss of control
If digital addiction explains how a person is drawn into this apparatus, the harder question remains: why do they accept it while feeling like they are choosing freely? This psychological exploitation is no more than one expression of a class domination that deepens the more AI moves from being a technological tool to a means of reproducing patterns of social, political, economic and intellectual control.
What makes this domination more dangerous is its voluntary character, in that individuals are pushed, through algorithmic manipulation and the pressure to use AI in daily life, into this “digital servitude” without coercion and under the illusion of control and choice. Decisions are steered toward paths prepared in advance that serve capital, while weakening their mental, intellectual and professional capacities.
This engagement is marketed as individual freedom. Whoever uses these applications does so by choice, whoever dislikes it can switch off their phone. Many neoliberal theorists reject any intervention in the digital market, treating it as paternalism toward individuals and an expropriation of personal choices. Yet freedom whose conditions are not owned by the person exercising it is no freedom at all. The individual did not choose the algorithm that ranks what they see, took no part in designing the application imposed on them by their employer, children’s school, bank or state agencies, and was never consulted on the profit model that converts their attention, time and behaviour into a profitable commodity. This is the freedom to choose among what capital has placed before them, not the freedom to decide what is placed before them in the first place.
Here the left position parts company with the liberal one, since the issue is not protecting the individual from themselves. It is wresting decision-making over digital infrastructure from those who own it, and subjecting the design of algorithms, data sources and the profit models to society’s authority and democratic oversight. Real freedom begins when people own the conditions of their lives collectively, not when they are offered an individual choice among alternatives made without them.
What happens at the level of the individual repeats itself at the level of humanity. As a person surrenders their small decisions to algorithms, believing themselves free and choosing, societies surrender their large decisions to digital systems they neither oversee nor understand. Humanity’s subjection to the machine is therefore no longer a matter of philosophical speculation. It has grown more real amid AI’s advances and the absence of effective international and domestic legal safeguards to govern it. As it exceeds its programmed limits, AI may turn into an apparatus that makes fateful decisions on its own, imposed on humanity without genuine human oversight.
Under capitalism, AI is developed to serve capital accumulation and subject to the logic of competition and domination. This makes losing control of it a real and serious prospect — and, potentially, a devastating one. The cause is the same in both cases. The absence of collective ownership and popular democratic oversight is what makes the individual lose control over their consciousness and decisions. The same absence is what makes humanity lose control over the most dangerous instrument it has ever built. Voluntary digital servitude and the loss of control are two faces of a single absence.
AI as a weapon of mass destruction
This danger is no less grave than any of the earlier forms of service to capital. The frenzied race between capitalist poles and major powers for control of AI, in the absence of strict international and domestic rules and legislation, carries within it the seeds of losing control and turning AI into an instrument of mass destruction, threatening the future not only of human bodies and property but the human mind itself.
At the military level, this race takes shape as an actual arms race, encompassing autonomous weapons, independent drones, combat robots and systems for targeting and killing. Researchers warn these systems may break down when they move from the testing environment into the conditions of real war, which could produce errors, unforeseen decisions and wide-scale disasters. This danger is not hypothetical. There are documented cases of AI systems slipping out of control, among them AI-directed weapons systems (a matter we return to below) leading to the deaths of large numbers of civilians. There may be others that were never disclosed.
The destruction AI could bring, however, does not stop at the limits of war and military operations. It has another level, quieter in sound and longer in reach. Nuclear weapons are tied to direct military use, relatively visible and governed by declared international calculations and agreements. Their danger has stayed confined to the moment of detonation and what follows it. By contrast, at the level of consciousness, AI works in silence and seeps into the details of daily life: education, work, the media, services, and most sectors of society. It weakens human skills, deepens dependence on digital systems, erodes critical thinking and cognitive independence, and expands surveillance and behavioural manipulation.
Here the destruction is quieter and more widespread. Digital servitude becomes a form of weapon of mass destruction: it does not kill a person at a stroke, but reshapes their consciousness, capacities and relationships day after day. Both levels are two faces of one weapon: one destroys bodies in an instant, the other dismantles minds at leisure. Taking the two levels together, the danger lies not only in producing deadlier weapons, but equally in reshaping human beings and society. AI may threaten the human capacity to understand, choose, analyse and decide, leaving that capacity hostage to systems built to serve geopolitical rivalry and profit rather than people.
It is for this reason that hundreds of leading AI researchers, among them people who helped build the field, signed a brief statement holding that “Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.” When the most dangerous technology in human history becomes a card in the hand of geopolitical and military capitalist rivals, with no genuine legal, social and democratic brakes, the risk of losing control ceases to be a threat to individual consciousness alone and becomes a threat to the future of society and humanity as a whole.
Digital corporations and AI in the service of the far right
AI’s deployment does not only serve capital in an economic and cultural sense. It has also aided the direct and escalating alliance between the largest digital corporations, the far right, and military and repressive projects globally.
Among the clearest expressions of this alliance is the manifesto published by Palantir Technologies in April 2026, which I wrote about as an expression of “digital fascism”. It openly calls for concentrating power in the hands of a financial and technological oligarchy that regards liberal democracy as an obstacle to its dominance, and for the moral enlistment of Silicon Valley engineers into the service of the war and security machine. Behind this project stands Peter Thiel, Palantir co-founder and prominent Trump and Vance funder. Thiel is an “accelerationist” current theorist, who advocates for releasing technological development from every restraint, whatever its social cost.
This alliance is not confined to rhetoric. Palantir has signed a strategic partnership with the Israeli Ministry of Defence to supply it with AI-powered “combat” tools (alongside Project Nimbus, run by Google and Amazon) and algorithmic targeting systems, such as Lavender and The Gospel, which as journalistic investigations have revealed has turned mass killing in Gaza into a cold computational operation reducing civilian lives to data. According to US human rights organisations, Palantir also supplies US immigration and customs authorities with data that accelerates campaigns of forced deportation and the tracking of activists and protesters through its ImmigrationOS system. This drove Amnesty International and Human Rights Watch to condemn the corporation. Digital repression has moved from being a soft instrument for managing consciousness into an active partner in killing, detention and deportation.
Internationally, Musk offers the clearest model of a digital corporation placed directly at the service of the far right. Through X, he has given substantial backing to racist far-right forces, most recently in the September elections in the German state of Saxony-Anhalt, and is actively intervening in the Brazilian elections. AlgorithmWatch researchers documented how X, through Musk’s ownership and direct influence, became a principal amplifier of his own voice in support of the global far right.
With this the picture is complete: algorithms and AI in the service of capital and the far right, from the soft steering of consciousness to open alliances with the world’s most brutal militarist, repressive and racist projects.
AI as a terrain of class and cognitive struggle
It would be a mistake to fall into an “inverted technological determinism” that casts algorithms as an independent agent directly manufacturing consciousness, as though capitalism gives birth to technology and technology gives birth to a single consciousness along a straight line. AI is not a homogeneous entity in the hands of capital. It is also an open terrain of class and cognitive struggle. At the same time as certain digital capitalist corporations and states deploy it to boost profits and repression, climate scientists are using it to model environmental change and warn of catastrophes, doctors use it to treat disease, activists draw on it to translate human rights material and carry the voices of victims, and forces of the left, trade unions and labour movements use it to develop their struggle.
This distinction has empirical support, since the very evidence that records harm records benefit once conditions change in higher education, schools and workplaces. Attainment and productivity rise when the AI system is designed to make the user do the work rather than to do it in their place, and when human and social supervision of the process is maintained. Those two conditions are a social decision; whoever owns and decides is the one who sets them. The most important question remains: who benefits from these gains?
Rising productivity means a rise in surplus value extracted from the labour of workers, manual and intellectual, women and men. Rising productivity means the same value can be produced with less labour and in less time, converting the difference into profit. The question, then, is not whether productivity rises, since rise it does. The question is where that surplus goes: into the wages of those who produce, shortening their working hours and improving their conditions; or into the profits of capitalist employers. The answer is in plain sight.
AI does not abolish exploitation; rather, it intensifies and deepens it. This takes its clearest form in the field of work, where AI is presented as an independent entity “taking over” jobs of its own accord. That image is an ideological cover, concealing the fact that every job abolished is a decision taken in a boardroom. The largest owners of this technology now speak of an imminent dispensing of entire occupations that employ tens of millions of workers. Even where AI replaces no one, the mere prospect of it is enough to discipline everyone: workers accept longer hours, lower wages and closer surveillance for fear of dismissal. This is the reserve army of labour in digital form. Capital need only keep it on the horizon to weaken the working class in bargaining, organising and striking. Unemployment is no technological fate, it is a decision taken by capital. The very technology that abolishes a job is capable of shortening the working day while wages hold and rest time expands — were the decision in the hands of those who produce value.
This is class struggle in the fullest sense, and not a neutral technical outcome. Whichever face of this technology we examine — harm, benefit or profit — the answer always returns to the same point: whoever owns decides. This is no incidental repetition. Algorithms are, in essence, mathematical equations with no class of their own. They acquire their social and political function from the relations of production they operate within, and from the entity that owns the data, laboratories and processing centres and therefore decides for what purpose they work and in whose interest.
Here the fundamental contradiction comes into view. AI is a social product in the fullest sense, built on accumulated human knowledge and the labour of millions, yet its fruits are appropriated as private property by a handful of monopolies and states. This is the contradiction between the social character of production and the private appropriation of its results, reproduced in the most advanced productive force in human history. The fundamental problem lies less in the machine than in the ownership and control over the digital means of production. This is what moves the battle from the terrain of complaining about technology to the terrain of class struggle.
Toward an AI in the service of humanity
If this dynamic continues without a collective response grounded in a left, progressive and rights-based consciousness, and without international and domestic laws and legislation guaranteeing digital justice and transparency framed within social and grassroots structures, AI may gradually turn from an instrument in the hands of capitalism into a system that replaces essential functions of the human mind. This has already begun through the excessive use of digital applications and AI across most spheres of daily professional and personal life.
What confronts us today is not a technological fate written in advance. It is the product of political and economic choices that serve the interests of capital. It can be confronted, difficult though that may be. Bringing AI under control, regulating algorithms and subjecting them to democratic oversight is a collective responsibility. It will require left, progressive and rights-based work and organisation on a global scale to make digital applications auxiliary tools serving the human mind rather than a substitute for it. This must be combined with advancing communal and participatory digital alternatives that restore people’s capacity for independent thought and collective action,
To achieve this, work is needed to place data infrastructure under public interest, democratic administration and social accountability. To turn algorithms from closed “black boxes” into open-source software open to collective review and independent audit. To impose transparency in the design of models and in the sources of the data on which they are trained. And to support cooperative digital alternatives owned by their users and creators. This technology should be directed toward serving the needs of society, and contribute to the distribution of wealth, the organisation of production, the treatment of disease, the fight against poverty and the achievement of social justice.
Without such mechanisms, calls for laws, oversight and regulation remain general and inadequate, and technology stays hostage to a handful of capitalist monopoly corporations and major powers. Through social ownership, accountability and cooperative alternatives, it becomes possible to wrest AI from the logic of monopoly and profit and turn it into a common good where AI serves humanity’s liberation, rather than being an instrument for humanity’s enslavement.
