Capitalism and Myth of AI Autonomy

Artificial Intelligence (AI) has become one of the most discussed and celebrated technologies of the 21st century, particularly among business magnates and tech enthusiasts. The rise of AI is often propagated as a herald of a new era, with some, like Elon Musk, even predicting that it could lead to the downfall of humanity by 2030. Similarly, Geoffrey Hinton, the ‘godfather of AI,’ has graced us with his apocalyptic musings, telling the BBC that there's a generous 10-20% chance AI might wipe out humanity within the next 30 years—because who doesn't love a touch of existential dread with their morning coffee?

Hinton’s apocalyptic warnings about AI leading to human extinction seem to overlook a fundamental condition of capitalism, the leading force behind automation: the surplus generated in production doesn’t just magically realise itself. It can only be realised with consumption for which humans are required. If AI wipes out humans, who will be left to purchase the goods and services produced by the capitalist system?

Consumption is not an isolated phenomenon in capitalism but is an essential part within the larger cycle of production, distribution, exchange, and consumption. Hence, it is dynamically related with the act of production and motivates the production by actualising its purpose, making not only the endpoint but also a catalyst for the production process.

Capitalism's internal dynamics also impose limits on the extent to which production will be automated. Capitalists, in their pursuit of maximisation of profit, strive to increase productivity by replacing workers with machines. As automation progresses, the proportion of living labour in production decreases relative to machinery, leading to a decline in the rate of profit in the longer run. The tendency of the rate of profit to fall restricts the economic viability of extensive automation in capitalism.

And while we're at it, who’s going to feed these ‘AI overlords’ their lifeblood—energy? In a world already battling an energy crisis, are we to imagine a dystopia where machines magically generate power, running on vibes and dystopian despair? The extinction scenario sounds more like the plot of a sci-fi film echoing Hollywood depictions of machines rising against their creators.

These dystopian depictions and narratives suggest that AI is not only a tool of human creation but also a force with autonomy and consciousness. It is even described as a ‘power-seeker’. The idea that AI could develop consciousness and even seek power is truly a ‘modern marvel’ -- a farce. Apparently, Darwin’s groundbreaking theory of evolution, which painstakingly traced the rise of human consciousness through millennia of labour, survival, and adaptation, hence, becomes redundant in the age of AI. The long and arduous process in which humans became conscious beings through their unity and struggle with nature—AI, with its pristine logic and impeccable algorithms, simply bypasses all that.

Consciousness is shaped by material conditions of the production of means of subsistence that shows that social existence determines consciousness. Human beings develop consciousness through a process of production for survival and social interaction, evolving within material realities. This means that consciousness is not a passive reflection of the world, as we are made to think, but a dynamic and active process. The notion that AI can develop consciousness ignores this scientific historical context.

Such portrayals mislead and advocate a thinking that AI has developed its own agency, separate from the social and economic structures that drive its creation and implementation. In contrast to these sensationalised interpretations of the mysterious power of AI and ‘mythologies of technological autonomy and machine intelligence’, Karl Marx provides a grounded framework to understand the roots and implications of automation.

For Marx, technological advancements are not mysterious forces arising independently; instead, these are products of the social relations and contradictions inherent in capitalism. The contradictions between labour and capital are the driving force for the continuous advancements in technology leading to the current phase of AI’s ‘almighty’ presence. The overall development of science and technology, including AI, is driven by capitalism's relentless pursuit of maximising profits.

The constant push to reduce the average time needed to produce a commodity (called socially necessary labour time), stems from the greed to increase the rate of surplus value which can be done at best by extending the time of labour performed beyond what is necessary to produce value equivalent to the workers’ reproduction (called relative surplus labour-time). The extension of the relative surplus labour time is not possible without contracting the socially necessary labour time. This dual requirement forces capitalists to innovate continually, leading to technological advancements that enhance productiveness of the labour. Thus, AI is a result of this condition, much like any other machine since the industrial revolution and not any mysterious force.

The replacement of living labour with machines is also being mythologised and framed in a very ahistorical manner. The retrenchment of workers due to automation and AI is often framed as a unique historical moment, a seemingly unprecedented phenomenon. Yet, such concerns are neither new nor groundbreaking. Hinton shares this worry about joblessness caused by AI. He even resigned from Google to speak freely about the risks, expressing that “a part of him now regrets his life’s work.” In response to the looming AI crisis, he proposes universal basic income as a solution to mitigate job loss.

Such understanding about job displacement due to technology is ahistorical in nature. The fear of machines replacing workers is not a new phenomenon but a recurring one, as the capitalist system has been relying on technological advancements in its relentless pursuit for profit.

In the early 19th century, after the Industrial Revolution, the Luddites smashed machines (mecahnised looms) which they saw as a threat to their jobs.  The propagation of projecting AI as a unique phenomenon overlooks this historical trajectory of capitalism. It also conceals the fact that the existence of a surplus population of workers serves as a disposable reserve army for capital, which means that the creation of a reserve army of workers is an integral part of capitalist development.

The presence of a reserve army of labour creates an environment in which the employed wage-labour always works under a constant pressure of losing the job and accepting low wages. This means that there is always a pool of workers available which can be exploited when capital needs to expand. For example, capital can draw upon this surplus population during the times of economic expansion without any increase in wages. Similarly, in the times of economic downturn, the profits can be maintained by relying on the surplus population without investing on any long-term labour benefits. This substitution of living labour with the machines signifies a change in the way surplus value is produced and wealth is accumulated, hence machines do not carry any mythical power, but the power of capital. 

Marx argued that as automation advances in production it reduces the workers to appendages of machines and they are relegated to the role of overseers, monitoring machines instead of directly engaged in production. It follows from this that accumulation of wealth at one pole is accumulation of misery and agony at the opposite pole.

While debunking the myth of AI's autonomy and mythological power, the aim here is not to critique technological advancements per se, but to highlight the socio-economic conditions under which these technologies are developed and deployed. Technological innovations, including AI, have undoubtedly increased human productivity. Many would argue that the labourer owes a debt of gratitude to capital for enhancing productivity by shortening the socially necessary labour-time i.e. the amount of labour-time required to produce a commodity under given social conditions of production at a given stage in the economic development of society or the average labour-time taken to produce a commodity.

But it must be noted that the shortening of the socially necessary labour-time is at the same time an increase in the relative surplus labour-time -- the portion of the day worker spends on producing beyond what is necessary for their means of subsistence. This surplus labour-time is the source of profit for the capitalist. This means that labour-time is central to the generation of profit and the capitalist system creates tensions around the length of the working day.

Labour-power is sold as a commodity, and its value is determined by the time required to produce the worker's subsistence. So, the tension is to increase the surplus labour-time and reduce the necessary labour-time so as to maximise the profit. This creates a conflict: the capitalist aims to extract as much surplus value as possible either by extending the working day or by reducing the necessary labour-time. This, in turn, leads to retrenchment of workers (as discussed above).

One of the growing concerns today is that skilled workforce in the service industry is being retrenched globally. The recent wave of layoffs in big tech companies, such as Meta, Amazon, Google, and X (formerly Twitter) etc., highlight how skilled workers, especially those in technology and software development, are increasingly being thrown out of jobs. Despite being highly skilled, these workers are replaced by AI-driven systems and automation tools that enhance productivity while reducing the need for human labour.

As capitalist production advances, it tends to generalise labour processes through mechanisation and automation. Complex tasks that once required skilled labour are broken into simpler components that machines or unskilled workers can perform. This deskilling reduces the value of labour power and its cost of reproduction, as it no longer requires extensive training or expertise, turning both skilled and unskilled workers into interchangeable cogs in the production process.

Technological advancement and mechanisation of the production process inevitably ends the categorisation and specialisation between the skilled and unskilled, as it generalises the labour. The wealth generated, therefore, continues to accumulate in the hands of a few giant corporates, while skilled workers, previously seen as irreplaceable, are now disposable in the relentless drive for higher profits. Understanding these dynamics helps expose how the benefits of technological progress are unevenly distributed, deepening existing social and economic divides.

The writer is Assistant Professor, Department of Sociology and Social Work, Christ University, Delhi. The views are personal.