Skip to main content

Full Marx What’s the future with AI?

The increased efficiency of any aspect of capitalism will not automatically lead to its collapse and transformation — avoiding the dangers and realising the potential of AI will involve struggle, argues the MARX MEMORIAL LIBRARY

PART ONE of this answer emphasised that information technologies, from “good old-fashioned artificial intelligence” of the 1960s and ’70s to robotics and machine learning today, greatly magnify the contradictions of capitalism.

They raise a host of questions: whether development should focus on medical diagnostics or personal surveillance; whether “we” need driverless cars and “just walk out” shopping and if so, whether they should be limited to those who can afford them or “owned” collectively; — and in either case, under whose control and scrutiny.

At a deeper level, it raises fundamental issues about the structural changes within global capitalism, the use of technologies for social control, repression — and war — and the required left and labour movement response.

Karl Marx in his early work on automation wrote of “the machine which possesses skill and strength in place of the worker; is itself the virtuoso, with a soul of its own in the mechanical laws acting through it.”

Popular images of AI tend to reflect its representation “in the movies” — often humanoid robots like the replicants in the 1982 film Blade Runner. In practice the closest analogy today is the character Holly, the on-board computer in the 1988 sci-fi space sitcom Red Dwarf — but without the humour.

From the earliest computer-generated imagery (CGI, as in Alfred Hitchcock’s 1958 film Vertigo) today’s augmented and virtual reality (AR and VR), “generative AI” — much of it now, like Metaverse, concentrated in a relatively few companies raises a host of ethical, legal and political, issues.

For example: “[I]f an AI system is observing hundreds of millions of human actions and replaying those actions through statistical correlations as ‘neural networks’ do, then who should be acknowledged as the author? Is it the programmers who harvested the data, or the workers (paid or unpaid) whose attention has been harvested?”

As the 2023 Hollywood Netflix writers’ strike emphasises, these are not merely abstract questions. Ominously, AI is now commonplace from consumer scams and “deepfake news” to military applications: AI lends an increasingly threatening meaning to “intelligence” services. As a recent feature article by this paper’s Science and Society team states: “Surveillance and data-gathering goes hand in hand with automation.”

But AI does not only consist of flows of information. Marx added, of his “automaton” that “it consumes coal, oil etc, [today it would be energy] just as the worker consumes food, to keep up its perpetual motion.”

AI requires capital, physical (the computers, and the human labour that makes and manages them) and financial.

One of the features of capitalism is the continued increase in the “organic composition of capital” (“OCC,” the ratio of “constant” capital  — plant, equipment and materials — to the “variable” capital of waged workers) and the consequent tendency for the rate of profit (as realised surplus value generated per unit capital, constant and variable) to fall. That tendency (the “TRPF’”) can be offset by many factors, not least the increased exploitation of human workers — see an earlier Q&A (number 104) on the OCC and TRPF.

But it is ever-present. For the capitalist, investment in AI itself increases the OCC though like other machine processes, it also offers the prospect of increased worker exploitation and control. Investment under capitalism takes place primarily for profit, not to raise output or productivity as such. When profit cannot be sufficiently raised through increasing what Marx called the “absolute” surplus value (longer hours, cutting rest breaks etc) then increasing “relative” surplus value (labour productivity) is the solution and technology — including AI — is increasingly the means. Moreover, the substitution of capital for labour applies in AI as in every other field of production, affecting IT professionals and other “intellectual labourers” alike.

In a hypothetical fully automated society, robots would do everything. Human labour would no longer be needed and would therefore by definition not be exploited. Productivity of goods and services would (similarly by definition) tend to infinity while profitability (surplus value extracted from workers) would tend to zero.

But if workers are not working and therefore no longer receiving any income, who would be able to buy? Who would the owners of the capital in this automated world be able to sell to in order to make a profit? Effectively, capitalism as the production of goods, for profit, extracted from the “surplus value” produced by workers, would collapse.

Scenarios like this underlie the support of people like Jeff Bezos, Bill Gates, Elon Musk, Mark Zuckerberg, and other “captains” of high-tech, for a universal basic income, paid for by taxation, to keep capitalism and profits afloat. They also underlie the arguments of texts like Aaron Bastani’s Fully Automated Luxury Communism and Paul Mason’s Post-Capitalism: A Guide to Our Future (both published in 2015).

The reality, of course, is that profits would decline and capital accumulation would be in trouble well before this. Some analysts are drawing parallels between the excitement over AI and the “dot-com” boom which collapsed in 2000. This isn’t just a matter of the bursting of a speculative bubble.

Marxist economist and blogger Michael Roberts declares: “The most important law of motion under capitalism, as Marx called it, would be in operation, namely the tendency for the rate of profit to fall.

“As ‘capital-biased’ technology increases, the organic composition of capital would also rise and thus labour would eventually create insufficient value to sustain profitability (ie surplus value relative to all costs of capital).

“We would never get to a robotic society; we would never get to a workless society — not under capitalism. Crises and social explosions would intervene well before that.”

Neither Marx nor Engels ever argued any more than Marxists today believe that capitalism will collapse “automatically” or transform itself into some less exploitative form. Capitalism has always shown a remarkable ability to accommodate challenges, albeit always at the expense of working people.

The challenge of information technology, old and new, is neither to anticipate a utopia nor to avoid a dystopia, though both have their limited place. Dystopian and utopian alternatives provide a useful focus for debate and speculation but should not divert from the more familiar and pressing struggles to build a force for socialism within workplaces and communities — and, of course, in Parliament.

Sir Tim Berners-Lee, inventor of the hypertext transfer protocol and the world wide web (yes, the http and www on every internet address) saw the web as essentially a liberatory device: he deplores the fact that it has fallen victim to the same laws of capitalism that have befallen every innovation, from electricity to train transport.

At the same time, within a sane, socialist society, AI has the potential to address issues which socialists have long pondered — not least, how to ensure that production to meet real consumption needs can be balanced. It has long been a claim by the right that only a “market society” (a misnomer if ever there was one, and something addressed in another Q&A) can ensure that supply and “demand” (the last itself often artificially created, not least in the so-called “fashion industry”) are balanced.

This was a problem recognised (for example) in Chile in 1971 under the newly elected socialist government of Salvador Allende, in its Cybersyn project. This aimed to link computers to a telex-based communication network in order to allow the government to regulate production in the state-owned sector of the Chilean economy while preserving the autonomy of workers and lower management. The fascist Pinochet coup in 1973 put an end to Cybersyn but AI opens up a whole new world of analogous possibilities within democratic socialism.

In China, AI is seen as a “new productive force,” an antidote to a declining “demographic dividend” (an ageing population) with applications from robotic ceramic glazing (immune from pneumoconiosis and working 24 hours a day) to television presenter avatars, broadcasting in different accents and languages.

In the meantime, the development of AI is driven by profit and power. AI amplifies the contradictions within a capitalist world economy, the burden of course being borne by the working class, and its applications reinforce the (racial, gendered as well as class) inequalities of its origins.

One consequence of capitalism, wrote Marx, is “estrangement” — “that finally — and this goes for the capitalists too — an inhuman power rules over everything.”

Capitalism will not “automatically” morph into some “post-capitalist” state (least of all, socialism). Ending capitalism will require a conscious, collective action on the part of “the many” — the working class.

Change will come when enough people see it as in their interests to secure it. Automation particularly AI, exacerbates the dysfunctionality of capitalism — and makes its replacement even more urgent.

As an earlier Q&A, on automation, declares, “It makes socialism necessary and communism possible.” Achieving them is up to us.

Monday April 29 sees the launch of Reds on the Green at 6.30pm — a programme of free events with an introduction to International Workers’ Day with labour historian Prof Mary Davis and trade unionist Alex Gordon, music and song from Raised Voices and a look at May Day through time. Details on the Marx Memorial Library’s website


We're a reader-owned co-operative, which means you can become part of the paper too by buying shares in the People’s Press Printing Society.



Become a supporter

Fighting fund

You've Raised:£ 7,865
We need:£ 10,145
14 Days remaining
Donate today