We are publishing here the report to the Ninth Congress of the Socialist Equality Party (US) given by Evan Blake. The Congress was held from August 2–7, 2026. This report introduced the fifth resolution, “AI and the fight for socialism,” which was adopted unanimously.

Comrade North’s opening report to this Congress assessed the first six years of the decade of socialist revolution. Among the fundamental tendencies it analyzed is the explosive contradiction between technological development and the profit system based on private ownership. That contradiction finds its sharpest expression in AI, the subject of Resolution 5. Never before has our movement devoted an entire congress resolution to a single technology. It has done so because, as the resolution states, in no other technology are the contradictions of the capitalist system so concentrated, and because the struggle over AI has already become a central battlefield of world politics and the class struggle.

This report elaborates the analysis underlying the resolution: the social character of this technology, the scientific revolution out of which it arose, the uses to which capitalism puts it, the significance of Socialism AI for the education of the working class, the speculative mania and the crisis of the profit system, the false solutions advanced by the pseudo-left, and the position the party must take.

An instrument of exploitation, surveillance and war

As the resolution states, “Under capitalism, AI is being introduced not to lighten the burden of labor but to cut it as a cost—the realization, for the corporations, of a long-cherished dream of drastically reducing the workforce.”

The scale is already vast. In just the first five months of 2026 there were nearly 400,000 job cuts announced, with AI the leading reason. Studies project the displacement of hundreds of millions of workers worldwide by 2030.

To review just a handful of the announcements of the corporations themselves, Amazon eliminated 30,000 corporate positions between October 2025 and January of this year, roughly 9 percent of its corporate workforce, while UPS has wiped out 48,000 jobs through the automation of its hubs. Microsoft cut some 15,000 positions over 2025, many of them software engineers, having announced that AI was writing as much as 30 percent of its code; Salesforce cut its support staff from 9,000 to 5,000, its chief executive explaining that with AI agents he needed “less heads.” The same process is underway on every continent, and is still in its earliest stages.

What distinguishes this wave of automation from all its predecessors is the character of the labor displaced. For the first time, automation is striking directly at intellectual and professional labor—coding, clerical work, translation, design, financial analysis—higher-paid layers of the working class whose jobs were long thought secure. For three-quarters of a century the promise of secure professional employment through education was a central mechanism of capitalist stability in America. That mechanism is now being dismantled. A social order that proletarianizes its educated youth and then denies them a future is manufacturing its own gravediggers, completing a process the Communist Manifesto foresaw, in which capital converts “the physician, the lawyer, the priest, the poet, the man of science” into its wage laborers.

In the same period the technology oligarchs have enriched themselves as never before, epitomized in the fascist Elon Musk becoming the world’s first trillionaire. In the process, the liberal veneer Silicon Valley once wore has been discarded. Musk finances fascist movements around the world and performs the Nazi salute openly, Google quietly dropped its motto “don’t be evil,” and the “Dark Enlightenment” ideologists whom the oligarchs cultivate propose quite explicitly that the American republic be replaced by a corporate monarchy under a chief executive.

Within this context, AI has been forged into an instrument of mass repression. The platforms collect the data of billions and willingly make it available to the state and the employers, while social media is scoured to identify students and workers for deportation over their political speech, above all their opposition to the Gaza genocide. Trump’s ICE fuses government databases, commercial records bought from data brokers, facial recognition and license-plate and location tracking into a single apparatus of coercion, built for it by Palantir. What is being assembled, piece by piece, is the technological infrastructure of a police state, whose target today is the immigrant and tomorrow the working class as a whole.

Above all, AI has become central to the new imperialist redivision of the world. The work of deciding who shall die is being transferred to software that generates kill lists at machine speed. In the US-Israeli assault on Iran, Palantir’s Maven Smart System—incorporating the Claude model built by Anthropic—produced some 1,000 bombing targets in the first 24 hours of the war, orchestrating the murder of thousands of civilians, including children. The same systems administer the genocide in Gaza, running on the Project Nimbus cloud infrastructure that Google and Amazon built for the Israeli state.

The ruling class does not conceal where this is heading. In July 2025 the White House issued its official strategy, “Winning the Race: America’s AI Action Plan,” whose opening line bluntly states the position of American imperialism, “The United States is in a race to achieve global dominance in artificial intelligence (AI).” The plan suspends environmental review to fast-track the data center buildout and organizes the export controls designed to strangle Chinese development, while ensuring that the frontier US laboratories are among the Pentagon’s major contractors.

As the resolution insists, “None of these reactionary uses flow from the technology itself, but from the social relations within which it is developed—private ownership, the profit system and the division of the world into rival nation-states.”

The Marxist position on technology

In response to these reactionary uses of AI, hostility to the technology is widespread and growing—among workers who see it destroying their livelihoods, artists whose work is plundered without payment, young people confronting a hollowed-out labor market and many rightly concerned by its environmental impact and use in waging imperialist wars. The socialist movement is by no means indifferent to these concerns, and the anger is understandable. Anti-AI sentiment is also not merely spontaneous, but being cultivated by every pseudo-left tendency, by sections of the far right and by broad layers of the political establishment, all of which have an interest in turning that anger against the technology rather than the class that owns it. Later in this report we will examine the pseudo-left’s response in more detail.

The conclusion that the technology itself is the enemy is a profound political error—a modern form of Luddism, named for the English textile workers who smashed the new machinery in the early 19th century. The Luddites were not fools; they were skilled workers defending their livelihoods with the only means available to a class that had not yet developed a consciousness of its historical position. To revive their standpoint two centuries later, after the entire experience of the international socialist movement, is not the inexperience of a young class but a conscious retreat from everything that experience has established.

The practical result of a categorical rejection of AI is not to disarm the corporations and the state, which will wield it regardless, but to disarm the working class alone. Luddism obliterates the distinction upon which the entire Marxist attitude toward technology rests, that between a machine and the social order in which it functions. The resolution anchors that distinction in a passage from Trotsky’s 1927 essay Culture and Socialism:

Technology is a basic conquest of mankind; although it has indeed served until now as an instrument of exploitation, it is at the same time the basic requirement for the liberation of the exploited. The machine strangles the wage-slave. But the wage-slave can only be freed through the machine.

Trotsky was addressing one of the central problems of the young Soviet republic, how a besieged and backward workers state was to raise the cultural level of the masses. His answer was bound up with the struggle against the theoreticians of “proletarian culture,” who held that the working class should discard the technique and culture of the old society as class-contaminated. Trotsky demolished this position, arguing that the proletariat had to master the entire inheritance of the society it had overthrown. He drew the conclusion with a directness that speaks across the century to the question before us:

We don’t have to shatter technology. The proletariat takes possession of the factories outfitted by the bourgeoisie, and it does so in the form in which the revolutionary overthrow found them… After all, the revolution was carried out precisely in order to seize this “heritage.”

What Trotsky grasped in the machinery of his own day holds for every conquest of nature humanity has made. The chemistry that learned to fix nitrogen from the air feeds nearly half of humanity, and filled the shells of two world wars; the physics that unlocked the nucleus incinerated two cities before it ever lit one. The science does not decide between abundance and annihilation. The social order decides, and until the working class takes power every such conquest will be turned first to profit and war. AI is the sharpest expression of this contradiction the world has yet seen, because it operates not upon matter but upon knowledge itself, upon the entire accumulated thought and culture of our species.

Advancing this Marxist approach towards technology, the resolution states that the working class “must take hold of AI technologies, as it must take hold of the factories, the mines, the data centers and the banks, and wrest them from the class that has monopolized them.”

The scientific revolution and the general intellect

Every revolutionary transformation of society has been prepared by a revolution in thought. But the revolution in thought is never the ultimate cause; it is itself the product of changes in the material foundations of society. The scientific revolution of the 17th century arose out of the growth of trade, navigation, mining and manufacture, which posed complex problems that had to be solved. In the work of Kepler, Galileo, Harvey and Descartes, and above all in Newton’s Principia of 1687, the authority of scripture and scholastic dogma over the understanding of nature was broken.

The Enlightenment of the 18th century carried the same method into society itself: Voltaire, Rousseau, Diderot, d’Alembert, Holbach and others subjected property, law, church and monarchy to rational examination, while Lavoisier placed chemistry upon a quantitative foundation. Engels described these men in Socialism: Utopian and Scientific:

The great men, who in France prepared men’s minds for the coming revolution, were themselves extreme revolutionists. They recognized no external authority of any kind whatever. Religion, natural science, society, political institutions—everything was subjected to the most unsparing criticism: everything must justify its existence before the judgment-seat of reason or give up existence.

Engels did not treat this intellectual upheaval as a self-generating movement of ideas. He explained it materialistically, establishing in the same work the standpoint from which every such development must be understood:

From this point of view, the final causes of all social changes and political revolutions are to be sought, not in men’s brains, not in men’s better insights into eternal truth and justice, but in changes in the modes of production and exchange. They are to be sought, not in the philosophy, but in the economics of each particular epoch.

The philosophers were themselves the product of a rising bourgeoisie and of the growth of the productive forces beneath the feudal order; their reason was the reason of a class whose mode of production had outgrown the political forms of absolutism. The intellectual revolution was the indispensable medium through which that class came to consciousness of its historical task. Diderot’s Encyclopédie assembled the learning of the age and placed it in the hands of that class, and the bourgeois-democratic revolutions in America and France, whose shock waves passed through Europe and across the Atlantic world from Haiti to South America, followed within a generation.

The modern scientific and technological revolution of the 20th and 21st centuries plays an analogous role today, on an incomparably higher level. It has assembled to the highest degree the material premises of a world socialist society, and it has produced an instrument capable of placing the accumulated theoretical and historical experience of the working class before the billions of workers who make up the world proletariat, at once and in every language.

Consider what has been achieved within the span of a single human lifetime. The structure of DNA was established by Watson, Crick and Franklin in 1953 and the genetic code cracked by 1966; the human genome was sequenced in 2003, and by 2012 the CRISPR technique had given humanity the power to edit it at will. In the same span the cosmic microwave background, detected by Penzias and Wilson in 1965, established that the universe itself has a history, and in 2015 the gravitational waves Einstein predicted a century earlier were directly measured.

Beneath all of it ran the revolution in computation and communication—the transistor in 1947, the integrated circuit in 1958, the microprocessor in 1971, the World Wide Web in 1991—until, within a single generation, more than half of humanity was joined in instantaneous communication and the working class of every continent bound into integrated chains of production.

The term “artificial intelligence” was coined at the Dartmouth workshop of 1956, but the neural-network approach that finally succeeded—pursued by Hinton, LeCun and Bengio against the current of their field—was scorned and starved of resources for decades, and was vindicated only when the internet’s oceans of data and the growth of computing power made the deep-learning breakthroughs of 2012 possible.

This entire development belongs to the historical epoch of imperialism. Almost all of it has unfolded since the founding of the Fourth International in 1938, on the eve of the Second World War, whose founding document characterized the period as the death agony of capitalism. Throughout this epoch, science has been fused with militarism. The Manhattan Project, mobilizing the finest physicists of the age during the war itself, produced not only the bombs dropped on two Japanese cities that were of no military significance but the instrument with which American imperialism intended to secure its postwar hegemony. Out of the same drive came the missile programs from which the space race grew, and the Pentagon communications network that later became the internet. Vast resources were placed at the disposal of research, and the price was the subordination of science to the war aims of the state. The distorting pressure of the imperialist epoch upon science is not a thing of the past; it is the medium in which AI is being developed.

The other pole of that epoch must also be stated. The October Revolution lifted a backward and largely illiterate country into the front rank of world science within a generation. This was accomplished in spite of the Stalinist bureaucracy, which physically annihilated a generation of the finest scientific and cultural cadre in the Great Terror and imposed upon biology the fraud of Lysenkoism, driving Nikolai Vavilov and many others to death in prison. Even so, Soviet scientists launched the first satellite and the first man into space, Kolmogorov placed probability theory upon rigorous foundations, and at the Institute of Cybernetics in Kyiv, Ivakhnenko and Lapa published in 1965 the first working learning algorithms for deep multilayer networks—anticipating by decades the deep-learning methods now dominant with modern AI systems. These were the achievements of nationalized property and economic planning, however grotesquely distorted by the bureaucracy, and they indicate what the working class can accomplish. 

In the field of medicine and public health, the 20th century saw the development of antibiotics, modern vaccines and the coordinated international campaigns that eradicated smallpox and drove polio to the threshold of extinction, and lifted life expectancy across the world. The sequencing of the genome opened the way to the treatment of disease at the molecular level, and within weeks of the appearance of a new pathogen its structure can now be read and a vaccine designed against it.

That capacity was demonstrated in the COVID-19 pandemic, when for the first time in the history of any pandemic, humanity possessed the scientific means to eliminate the virus, as proven in China, New Zealand and other countries. But the vast majority of capitalist governments refused to use these measures, rejected elimination, and let the disease rip through the population, because the flow of surplus value could not be interrupted. Over 30 million people have now been killed by COVID and hundreds of millions more debilitated by Long COVID. The subsequent official embrace of anti-scientific reaction—vaccine opponents elevated to the direction of public health, the institutions of disease control dismantled—represents the wholesale repudiation of the scientific inheritance of the 20th century. 

The earth sciences establish the point with equal force in the case of climate change. Continuous measurement of atmospheric carbon dioxide began in 1958, and humanity has watched the concentration climb, year upon year, for 68 years. Carl Sagan explained the greenhouse effect to the US Senate in December 1985; James Hansen’s testimony followed in 1988, the year the IPCC was founded. The sources of global warming have been known for the entire lifetime of most comrades at this Congress, and every one could be eliminated with existing technology. What has been lacking is not knowledge but a social order capable of acting upon it.

Two conclusions follow. The first is the pace: the whole edifice of modern science, from the structure of matter to the code of life, has been erected within a single human lifetime. The second is its social character. Names still attach to these discoveries, but ever more thinly. The human genome was sequenced by hundreds of laboratories on four continents, and the detection of gravitational waves was reported over the signatures of more than a thousand scientists. Science ceased to be the work of isolated individuals and became a vast collective enterprise—organized, industrialized and international—and in the process became what Marx had foreseen a century before: a direct force of production, the general intellect of humanity materialized in social labor. 

The resolution places AI as the culmination of this entire historical development. It is the point at which the accumulated knowledge of the species is materialized in a machine, a social process which Marx identified in the early stages of industrialization in the Grundrisse:

The development of fixed capital indicates to what degree general social knowledge has become a direct force of production, and to what degree, hence, the conditions of the process of social life itself have come under the control of the general intellect and been transformed in accordance with it.

In Marx’s day this tendency could be glimpsed in the embodiment of science in machinery. In ours it has been consummated. These systems are trained upon the recorded thought of humanity—its literature, its science, the labor of every researcher and of every anonymous worker who wrote a manual or answered a question online. Capitalism began with the enclosure of the common lands; it approaches its end with the enclosure of the intellectual commons of the entire species.

Socialism AI and the development of socialist consciousness

The party’s approach to this technology has emphasized a terminological clarification, and it is not a quarrel over words. Marx established, in his analysis of the fetishism of the commodity, that under capitalism the products of human labor confront their producers as independent powers, and that the social relation between producers appears to them as a relation between things. The term “artificial intelligence” is a modern form of this fetishism. The word “artificial” severs this intelligence from the human labor that produced it and denies its authorship, presenting the collective intellectual inheritance of the species, appropriated and set in motion by a handful of corporations, as a mind that has arisen on its own and now stands above society. In an essay published December 7, 2025, days before the launch of Socialism AI, Comrade North wrote:

In reality, what is called “AI” is better understood as augmented intelligence—an extension and amplification of human intellectual labor. Its foundations lie in centuries of accumulated human practice and knowledge. … The algorithms do not invent meaning; they learn from vast human-generated corpora. … Their failures reveal the limits of human training, not the existence of some alien mind. The intelligence is not artificial; the automation is.

Five days later, the ICFI launched Socialism AI, conceived, built and maintained by our cadre. In doing so, we placed the heritage of Marxism—the works of Marx, Engels, Lenin, Trotsky, Luxemburg and Plekhanov, and the archives of nearly three decades of the World Socialist Web Site—at the disposal of workers and young people entering into struggle everywhere, at once, in any language, at any hour.

For the first time, a revolutionary organization placed into the hands of the working class a system able to answer serious questions of theory, history and organization from a Marxist standpoint, with citations to the sources. The statement published to coincide with the launch explained:

The decisive significance of Socialism AI lies in its ability to bridge the gap between the objective movement of the working class and the subjective level of socialist consciousness. Marxism has always insisted that the spontaneous struggles of workers, however powerful, do not by themselves produce a coherent revolutionary perspective. Consciousness must be developed; historical experience must be assimilated; theoretical insight must be achieved. The technological developments of the present era now permit the rapid transmission of ideas on a global scale, enabling the working class to develop its political understanding at a pace suited to the objective crisis.

This is the central strategic problem of the epoch. The objective crisis of capitalism is maturing at breakneck speed, while the consciousness of the working class has been thrown back by a century of Stalinist and social-democratic betrayals. The entire history of the Fourth International is the struggle to resolve the crisis of revolutionary leadership, and the International Committee has grasped that the very technological revolution capitalism turns against the working class provides an unprecedented weapon in that struggle.

Socialism AI is a weapon of revolutionary leadership, not a substitute for it. No technology builds a party or makes a revolution on its own, but only through the conscious activity of living revolutionists organized in the ICFI and its sections. But the weapons wielded by a revolutionary movement shape what it can accomplish. Diderot’s Encyclopédie placed the learning of an age at the disposal of a rising class; Lenin’s Iskra forged a party out of scattered circles across an empire; the World Socialist Web Site, launched in 1998 in a farsighted response to the emergence of the internet, made the analysis of the Fourth International available daily on every continent. Socialism AI stands in that lineage, and in the months since its release it has been used by thousands of workers and young people in more than 100 countries.

Socialism AI is a powerful force multiplier for the whole educational work of the party as we fight to build in the working class. It is the revolutionary instrument through which workers and youth can find answers to every political question they have, through immediate, interactive engagement. A worker who has not yet read Marx can begin with his own experiences—with the betrayals in his own plant, the war his government is waging, the social conditions he confronts—and be led from that question into the whole body of Marxist analysis. The millions being drawn to socialist politics by the deepening war and social crisis must be made aware of Socialism AI, which meets their objective needs for political education and training.

The resolution accordingly commits the party to make Socialism AI known as widely as possible and to bring it into the daily political activity of workers and youth, and this must be translated into definite practices. Comrades should demonstrate Socialism AI directly—at campaign tables, at party and IYSSE meetings, in workplace discussions—rather than merely describing it, for the decisive experience is that of a worker posing his or her own question and reading the answer. In the rank-and-file committees it should become a working tool, used to research an employer’s finances and global operations and to draft and sharpen statements.

Comrades should also make it part of their own writing and research, which means not asking it to write in our place but using it across successive drafts to gather and check material, sharpen an outline and test a formulation, with the comrade exercising political judgment at every step. It should be used in tandem with other AI tools, which are developing rapidly and can supplement any gaps, as we work to continuously improve and expand the capabilities of Socialism AI.

The discussion at this Congress on Mehring Books and the importance of reading the literature of our movement was critical, and I want to urge comrades to use Socialism AI systematically as a reading assistant, submitting any questions you have that arise in the course of studying the foundational texts of our movement. We are developing tools to integrate the Marxist Library on the World Socialist Web Site directly with Socialism AI, allowing comrades and all users to highlight and bookmark passages and to interrogate them as they read, which we hope to complete in the coming months.

Every branch and every member must take up the fight to promote Socialism AI as widely as possible, and in the weeks ahead we will be developing a campaign focused on this which will aim to win thousands of new users in the months ahead. This will require the active involvement of every member, including through promotion across social media, systematic contact work, and more. The new version released to coincide with May Day produced an important rise in usage and in the creation of accounts, and those who came stayed and asked further questions.

The integration of Socialism AI into the front page of the World Socialist Web Site and into every article, carried out last week, joins the system to the whole of our published material and demonstrates its value to every reader. This will soon be extended to all the language sites, and we will work to build new integrations of Socialism AI with the WSWS in the months ahead. Masses of workers and young people are today calling themselves socialists. They will be drawn to a revolutionary application of this technology, and every comrade is responsible for bringing it to them.

The AI bubble and the declining rate of profit

In the final part of this report, we will turn from the party’s own use of AI to the broader significance of the technology itself under capitalism and its potential use in a socialist society. The AI bubble is a speculative mania without precedent in the history of capitalism. As the resolution notes, a small group of interlinked corporations have bound themselves together in a web of “circular deals” that inflate the appearance of demand—Nvidia investing in OpenAI, which spends the investment on Nvidia chips, while Oracle books hundreds of billions in cloud contracts to serve a customer with no immediate prospect of profit. Such commitments now approach $1 trillion.

Nothing illustrates the character of the mania more sharply than the flotation of SpaceX on June 12, the largest public offering in history, which raised some $75 billion, valued the company at nearly $1.8 trillion and made Musk the world’s first trillionaire in a single afternoon. Central to the prospectus was the proposal, following the absorption of xAI, to launch a constellation of a million satellites to serve as data centers in orbit. Whether such a thing can even be built is disputed by engineers and by Musk’s own competitors, one of whom pronounced the idea ridiculous. For the financial oligarchy, it hardly matters. Under conditions in which the terrestrial buildout is running into the limits of land, water and electric power, a scheme of sufficient grandeur, advanced by a figure of sufficient notoriety, is worth hundreds of billions of dollars whether or not it ever functions.

By the economist Jason Furman’s calculation, investment in information-processing equipment and software was responsible for 92 percent of US economic growth in the first half of 2025, which means that outside the AI bubble the American economy is essentially stagnant. The fragility of the structure has already been demonstrated. In January 2025, the release by the Chinese laboratory DeepSeek of a model trained at a fraction of the advertised American cost erased nearly $600 billion from the market value of Nvidia in a single session.

The Chinese laboratories have gone on releasing capable models, many open-weight and far cheaper than their American rivals—a question not of technological breakthrough but of ownership and access, which exposes as illusory any permanent American monopoly over this technology. As with the railway mania of the 19th century and the dot-com bubble, a genuine revolution in the productive forces has become the occasion for a speculative frenzy that must collapse.

The deepest significance of the frenzy lies elsewhere, which the resolution summarizes: “This mountain of speculation rests on the promise that AI will restore the profitability of a stagnating economy. It will do the opposite.” Drawing upon the critical analysis made by Comrade North in his lectures in Berlin and London last autumn, the resolution states:

Value, as Marx demonstrated, is created by living human labor, and profit is derived from the surplus value extracted from the worker over and above the wage paid for his labor power. As capitalism develops, the mass of constant capital—machinery, plant and raw materials, and now the vast infrastructure of chips and data centers—grows relative to the variable capital laid out in wages, and the rate of profit tends to fall. By displacing living labor, the sole source of surplus value, on a vast scale, AI accelerates this decline. Far from rescuing the profit system, it drives to a new extreme its fundamental law of the tendency of the rate of profit to fall.

Why, in defiance of this logic, does the bubble continue to swell? Partly because of the ideology that now grips Silicon Valley—an accelerationism hardened into quasi-religious faith, whose central article is the imminent arrival of “artificial general intelligence” (AGI), a machine capacity surpassing the human in every domain which will supposedly usher in an age of “abundance.” Whether or not systems of such generality are eventually built, the claim that their abundance would flow to all stands reality on its head. It is essentially warmed-over trickle-down economics: the assurance, repeated by every apologist for capitalist accumulation since the 1980s, that wealth heaped up at the summit will somehow find its way down to those whose labor produced it. But under the existing property relations, every gain in productivity is appropriated by the owners, and the promised abundance materializes as staggering wealth at one pole and mounting poverty at the other. The AI bubble is only accelerating these laws of capitalist development identified by Marx over 150 years ago.

The AI economy and whatever benefits it yields will be captured by the top 10 percent, the ruling class proper and the affluent upper-middle layers, deeply hostile to the working class, who are living in the most literal sense inside the bubble. Their salaries, stock portfolios, retirement accounts and property values are staked upon the same speculative structure, which is why they experience the boom as general prosperity and cannot conceive of its collapse. 

The development of AI has at the same time confirmed, more nakedly than any previous development, a truth long established by the socialist movement: those who own the technology corporations and appropriate the profits contribute nothing to the creation of the value they appropriate. The resolution states:

The AI boom exposes the parasitic and superfluous character of the capitalist class. The owners contribute nothing to production. These systems are conceived, built, trained and maintained by workers—the engineers and researchers of Silicon Valley and its counterparts across the world; the low-paid data workers who label their outputs in Kenya, the Philippines, India and Venezuela; the miners of cobalt, lithium and rare earths in the Congo and other oppressed countries. … Society has no need for the oligarchs: their grip on economic life has become incompatible with social progress and the survival of civilization.

From the superfluousness of the owners the resolution draws the decisive conclusion:

The struggle over AI poses point-blank the question of who is to control the productive forces, and demolishes the oldest argument against socialism—that a modern economy is too complex to be consciously planned, that without the market and the guiding hand of owners and managers production would collapse into chaos. This was always a self-serving lie; with AI and modern communications it has become an absurdity. Workers organized in rank-and-file committees can now obtain in moments the answer to any question about their industry, their supply chains and the world economy, and coordinate instantly with their class brothers and sisters across every border. The material premises for the socialist planning of the world economy are fully in hand.

Augmented intelligence in the service of humanity

The liberating potential of augmented intelligence is already visible, in distorted form, within the confines of the profit system. Directed by scientists, AI has predicted the structure of nearly every known protein—a decades-old problem in biology, recognized with the 2024 Nobel Prize in Chemistry—and, under the direction of mathematicians, made contributions to open problems in pure mathematics. These are not the feats of an autonomous machine mind but of human beings wielding a powerful new tool, and they indicate what could be accomplished when freed from the drive for profit.

The environmental objection to AI must be answered concretely. The concern is legitimate in that the data center buildout under capitalism consumes electricity and water on a colossal scale, driving up ordinary people’s bills, reviving coal plants, and contributing to different forms of pollution. But the waste is a property of the profit system, not of the technology. A half-dozen corporations are racing one another to build duplicate models and duplicate data centers, each training its own system on much the same corpus of human knowledge.

Under a centralized, rational plan that duplication would simply disappear, and with it a vast share of the energy now consumed. Furthermore, the same technology under social ownership would become the most powerful instrument yet devised for solving the climate crisis it is presently made to deepen. Data centers are sited today according to the availability of tax breaks, cheap land and deregulated power, and often fed by reactivated coal or new gas power plants. Solar, wind and battery storage, deployed according to a rational plan, could power a network of net-zero data centers and begin rebuilding the world’s infrastructure on a carbon-free basis. Capitalism has not done this and never will, because it subordinates every such decision to private profit.

Under a democratically planned world economy, humanity would be able to tackle the greatest planning problem we have ever confronted, the transition to a carbon-free economy. The coordinated reconstruction, within a generation, of the world’s energy, transport, industrial and agricultural systems poses precisely the types of highly complex problems in which AI systems excel: balancing continental power grids built on variable solar and wind, designing better batteries, modeling fusion plasmas, detecting from orbit the methane leaks and deforestation that now go unmonitored, and more.

The energy question itself would be posed on an entirely different basis. Fission already generates vast quantities of carbon-free electricity, though under private ownership it has been developed erratically and saddled with the profit motive’s indifference to safety and waste. Fusion—the reaction that powers the sun—promises energy that is genuinely clean and effectively limitless, drawn from fuel in seawater, with neither long-lived radioactive waste nor any possibility of a runaway reaction; it is no longer science fiction but an engineering problem on which real progress is now being made. In a rationally planned world economy these forces, together with solar, wind and storage, could put an end to energy scarcity as a limit upon human development. Again, the obstacle is neither knowledge nor technique but the social order.

Against the false solutions of the pseudo-left

Everything therefore depends upon the political program with which the working class enters this struggle, and it is here that the pseudo-left performs its assigned role. As mass layoffs turn public opinion against the technology monopolies, the ruling class requires mechanisms to contain the anger. The Democratic Socialists of America and its house organ, Jacobin, have rushed to supply them, making AI a central preoccupation over the past year and promoting the neo-Luddite position.

It began with the calls for moratoria on data-center construction, to which Bernie Sanders and Alexandria Ocasio-Cortez gave legislative form in March through the AI Data Center Moratorium Act. Jacobin celebrated this in an article headlined “Stop the AI Build-Out, Start the Fight,” approvingly describing its “nonpartisan” coalition of suburban homeowners and “conservative farmers” and declaring that “the Left should be uniformly on its side.” What is proposed is an alliance with property owners and the political right against data center construction, leaving ownership of the technology in the hands of the ruling class and doing nothing to halt its use by the Pentagon or by ICE. The DSA claims the campaign as a success, reporting that nearly half the data centers planned this year have been delayed or canceled.

The Pabloite tendencies have gone even further. In May of this year, International Viewpoint, organ of the Pabloite United Secretariat, published “Theses for an ecosocialist critique of artificial intelligence” by Daniel Tanuro, which declare that the answer to AI is “the ecosocialist degrowth of global material production” and that the prospect of planned abundance belongs to “magic” rather than reason. The problem is no longer the capitalist relations that imprison the productive forces but the productive forces themselves. In a world where billions lack adequate housing, medicine and electricity, “degrowth” is a program of scarcity and rationing imposed on the working class.

The debate turned in June, when Sanders proposed legislation to establish an “AI sovereign wealth fund” in which the federal government would take a 50 percent equity stake in the largest AI corporations and supposedly pay the population a dividend. It was advertised as a blow against the billionaires, but converges entirely with the policy of Trump, who has already moved to establish a US sovereign wealth fund. It aligns as well with the wishes of the AI executives themselves, Sam Altman foremost among them, who have offered the government a stake in their firms with an eye to a bailout when the AI bubble inevitably bursts.

Sanders’ proposal set off a public debate in the pages of Jacobin, which published at least eight articles on this topic and AI more broadly in the span of a fortnight. Three distinct positions emerged.

The first fully embraced Sanders’ nationalist and pro-capitalist position. Dustin Guastella, an official of Teamsters Local 623, opened the series on July 9 with “The Case for Nationalizing Artificial Intelligence,” resting his claim on the fact that these systems were trained on the collective work of humanity. He wrote, “we have a distinctly patriotic case” to nationalize the AI companies, because public ownership of AI “is in our national interest.” The collective labor of the world’s workers is thus converted into a claim on behalf of the American nation-state, in the very race driving Washington toward war with China.

The second position answered that ownership is not control. Michael A. McCarthy, in “Don’t Just Nationalize AI. Democratize It.,” argued on July 16 that a state-owned AI industry would remain in the hands of the same state that runs the deportations and the wars, and proposed instead that decisions over AI be handed to a popular assembly chosen “via a random selection (or sortition).” The working class does not appear in this scheme as a social force at all, only as a pool of individuals from which names may be drawn, and the foundation of capitalist property is left wholly untouched.

The third position, styling itself the most radical, rejected nationalization from the “left.” Academics Hagen Blix and Ingeborg Glimmer wrote in “The Socialist Case Against Nationalizing AI” that AI is above all “a weapon of class war from above,” which no change of ownership can redeem, and called for “building a socialist movement against AI.” This is the Luddite conclusion in its purest form, and it arrives, significantly, dressed as the most “radical” of the three.

Across all their disagreements these tendencies share a single premise: not one raises the expropriation of the technology monopolies by the working class, the conquest of political power or the socialist reorganization of the world economy. Every one of these proposals ends in the capitalist state—as owner, as regulator, as chartering authority, as convener of assemblies. Under all of them, as our resolution states, “the corporations remain private and profit-driven, while the working class owns nothing, controls nothing and decides nothing.”

The coming social conflict and the tasks of the party

The ideologists of the tech industry present the outcome of the struggle over AI as already settled. Whether they promise abundance or warn of catastrophe, they proceed from the same premise: that the technology will itself decide the fate of humanity, and that the working class figures only as a casualty. This is fatalism, and it abstracts the technology from the class struggle. Between the present order and any reorganization of society stands the working class, which will not accept in silence the destruction of the livelihoods of hundreds of millions. The struggle over the control of AI will enter the strike movement now sweeping every continent. The working class must advance its own independent program: the expropriation of the monopolies and the conquest of political power.

That program is stated in the three demands the resolution advances.

• Expropriate the technology monopolies! The great AI, cloud-computing, semiconductor and digital-platform corporations must be expropriated and transformed into public utilities under the democratic control of the working class.
• Not a single job lost to AI! For a shorter working week and workers’ control over technology! The immense gains in productivity must be used to shorten the working week with no loss of pay, and the introduction of new technology placed under workers’ control.
• Turn AI to human need and the planning of the world economy! Under social ownership and a democratically planned world economy, AI must be turned toward the needs of humanity—the advance of science, medicine and education, the abolition of poverty and drudgery, the solution of the ecological crisis and more.

Nothing in this is automatic. Everything turns, in the end, on the conscious struggle of the working class and the resolution of the crisis of revolutionary leadership—on the work of this party. The resolution states the perspective in its closing sentences:

The working class must take hold of AI, as it must take hold of all the productive forces its labor has created, expropriate the oligarchy that has usurped them, and reorganize the world economy on socialist foundations. This is the perspective for which the Socialist Equality Party fights.

Comrade North’s opening report has situated the work of this Congress within the fifth phase in the history of the Trotskyist movement, which opened in 2019 and which the summer school of that year defined by the further colossal development of globalization combined with new technologies that have internationalized the class struggle to a degree hardly imaginable a quarter-century earlier. The emergence of AI at the center of world economy and world politics in the seven years since is the sharpest confirmation of that assessment.

But nothing in this permits an objectivist complacency. The crisis produces no revolutionary conclusion of its own accord. As Comrade North put it, the analysis before this Congress is not a commentary upon events but a guide for the party’s intervention in them. The expansion of Socialism AI, the building of the rank-and-file committees, the growth of the IYSSE, and above all the building of the Socialist Equality Party—these are the tasks through which the objective crisis is transformed into a conscious revolutionary movement.

Comrades, the purpose of this resolution is to arm the cadre for the period now opening: to enter the coming battles understanding this revolutionary technology more profoundly than the class that owns it, and to fight on that basis to place it, with all the productive forces, in the hands of the class whose labor created it. I urge the Congress to adopt the resolution.