Antonio Garcia provides a Marxist perspective and programmatic response on AI. This is Part 1 of 2. You can read Part 2 here.
The Presence of the Question
Artificial intelligence (AI) has been an unavoidable question in recent years. Ever since the first surreal scribbles of Crayon appeared on our screens, the technology and its discontents have been impossible to miss. This is particularly true among leftist and socialist circles. Every day, thousands of online leftists rail against ‘AI slop,’ against the ‘enshittification’ of social media, and against the job-destroying capabilities of the technology. Each of these are valid critiques. It is doubtlessly true that the vast majority of content produced by AI is, at best, unremarkable ‘slop,’ and, at worst, actively used to spread disinformation. It is true that the technology has greatly simplified content creation at the cost of personal input and creativity. It is true that certain jobs are becoming redundant in increasing numbers as artificial intelligence advances from a gimmick to an integral element of global capitalism.
This said, many of the critiques and practical ends put forth by leftists to ‘resolve’ the question of artificial intelligence have run counter to Marxism. Marxism is foremost a materialist science; it is one interested in material reality, material responses to reality, and understanding historical progression- better known as historical materialism. The leftist response to AI has, in contrast, been resolutely reactionary, activist, and idealist in nature. On the one hand, legions of online leftists seek to ‘name and shame’ AI out of the social sphere, launching into lectures, witch-hunts, and campaigns to dissuade individuals and corporations from utilising AI. On the other, there is a push, once again almost-entirely taking place online, which seeks to abolish the technology outright. These ‘campaigns’ are driven by a variety of motives: moral qualms about ‘the spirit of art’ being lost, solidarity with those whose jobs are at-risk of being replaced, environmental concerns, and the defence of property which AI devalues. Yet in doing so, these leftists abandon the scientific lens Marxism provides, and with it, any ability to comprehend and affect the development of society.
It is then apparent that the leftist common sense which currently guides discussion around AI is unsatisfactory for a sober, Marxist analysis of the issue. In this article, I shall give a brief overview of Marxist responses to various facets of the AI debate. Then, with a correct frame of mind established, I shall pass over the debate to you, the readers of Partisan! to discuss further the matters of technology and Marxist strategy.Confronting such an important-yet-emerging facet of global capitalism will require a clear mind and firm organisational unity. Thus, whilst I will lay out the framework of a Marxist approach to AI, the resolution of the question will require a broader consensus.
Marxism: Against Reaction, Activism, and Moralism
Our foremost concern with AI is how the debate surrounding it interacts with Marxist theory. To construct a programmatic response to the issue, we must first comprehend it within the context of Marxism. For the sake of structure, I shall centre our literature review around leftist responses to AI. From this left-liberal focal point, we will be able to unearth and critique the failings of the modern ‘movement’ and subsequently course correct.
The Reactionary Assertions
The first object of our study shall be the modern left’s demand to reject artificial intelligence entirely. To be clear, this section will focus on AI at a societal scale, not the rejection of AI on an individual or community basis. Of all the claims put forth by the modern left, the call for the wholesale rejection of AI is perhaps the most reactionary, as well as the most hopeless. However, hidden within this flawed dogma lays a kernel of truth which may be of utility to us. To understand this kernel, we must understand the basis of leftists’ ire for AI. On the societal level, we find this to be a matter of job preservation. Artificial intelligence is out-competing humans in various fields: data entry, art, customer service etc., and therefore must be rejected to preserve these jobs. Unfortunately, these critiques too must be subdivided. On the one hand, we have the genuine concern of workers who fear the loss of their jobs, whilst on the other we find significant push-back coming from the petite-bourgeoisie who mask their fear of being proletarianised under the same pretext.
What was Marx’s approach to such fears? Fortunately, he was abundantly clear on the matter. I quote from Capital:
“The instrument of labour, when it takes the form of a machine, immediately becomes a competitor of the workman himself. The self-expansion of capital by means of machinery is thenceforward directly proportional to the number of the workpeople, whose means of livelihood have been destroyed by that machinery. The whole system of capitalist production is based on the fact that the workman sells his labour-power as a commodity. Division of labour specialises this labour-power, by reducing it to skill in handling a particular tool. So soon as the handling of this tool becomes the work of a machine, then, with the use-value, the exchange-value too, of the workman’s labour-power vanishes; the workman becomes unsaleable, like paper money thrown out of currency by legal enactment.”
[…]
“But machinery not only acts as a competitor who gets the better of the workman, and is constantly on the point of making him superfluous. It is also a power inimical to him, and as such capital proclaims it from the roof tops and as such makes use of it. It is the most powerful weapon for repressing strikes, those periodical revolts of the working-class against the autocracy of capital. According to Gaskell, the steam-engine was from the very first an antagonist of human power, an antagonist that enabled the capitalist to tread underfoot the growing claims of the workmen, who threatened the newly born factory system with a crisis.”
[…]
“Although machinery necessarily throws men out of work in those industries into which it is introduced, yet it may, notwithstanding this, bring about an increase of employment in other industries. […] Since every article produced by a machine is cheaper than a similar article produced by hand, we deduce the following infallible law: If the total quantity of the article produced by machinery, be equal to the total quantity of the article previously produced by a handicraft or by manufacture, and now made by machinery, then the total labour expended is diminished. The new labour spent on the instruments of labour, on the machinery, on the coal, and so on, must necessarily be less than the labour displaced by the use of the machinery; otherwise the product of the machine would be as dear, or dearer, than the product of the manual labour. But, as a matter of fact, the total quantity of the article produced by machinery with a diminished number of workmen, instead of remaining equal to, by far exceeds the total quantity of the hand-made article that has been displaced. Suppose that 400,000 yards of cloth have been produced on power-looms by fewer weavers than could weave 100,000 yards by hand. In the quadrupled product there lies four times as much raw material. Hence the production of raw material must be quadrupled.”
What does this mean? To summarise, the innovation of new technologies creates competition between humans and machines, one that almost always results in machines’ victory. Machines, owing to their superior productive capabilities, then render entire fields obsolete by nature of the lower labour power required to furnish them, and thereby, lower costs.[1] This also allows machines to serve as strike-breakers, both by lowering the cost of labour power, thereby lowering the demands of workers, as well as by serving as unrepentant scabs should workers go on strike. Yet despite the perceived tendency to lower the amount of labour power in the economy, machines cause the inverse. Though they stamp out careers in their field, the increased productivity they cause and the demand for the machines themselves drives workers and capital into markets which supply them.
We therefore find ourselves in an uncomfortable position. Whether we use Marx or not, it is obvious that AI is displacing workers within the fields it penetrates. Similarly, we do not need Marx’s historical examples to understand how AI can be used to crush strikes. Yet what we find in Marx is not a condemnation of machinery, but instead, what appears to be the opposite. Indeed, he noted that,
“It took both time and experience before the workpeople learnt to distinguish between machinery and its employment by capital, and to direct their attacks, not against the material instruments of production, but against the mode in which they are used.”
In other words, he argued that the recognition that it is capitalism, not machinery, which alienates the worker from their labour and exploits their toil represents a maturation in the workers’ movement. Moreover, Marx understood that the loss of jobs by machinery is outstripped by the surge in jobs brought about in adjoining fields. Thus, whilst Marx acknowledged the ways machinery intensified the exploitation of workers within certain fields, he also noted that they brought about improved conditions for workers in adjacent fields, and that directing attacks against capitalism, rather than machinery, is an advancement in working class consciousness.
The very conditions outlined by Marx exist today. Service and finance-centred economies are experiencing job loss as AI invalidates the labour power of millions. In Australia, unions report that one third of jobs are at-risk of being eliminated by AI. Data from the United States’ Federal Reserve suggests that the adoption of AI will raise unemployment, a problem exacerbated by America’s already-existing joblessness epidemic. The situation in the United Kingdom is so severe that even Bank of England’s governor himself has sounded alarms over AI’s displacement of workers.
At the same time, economies predicated on the production of AI hardware have seen explosive growth. The government of Taiwan has launched several initiatives to onshore the production of AI systems, aiming to create 500,000 jobs in the process. AI has burdened China’s job market, not because workers are being laid off en masse, but because there are three jobs in the AI sector for every one qualified worker. Malaysia perhaps most clearly demonstrates this dichotomy: a recent study by the Ministry of Human Resources found that 620,000 jobs are likely to be lost to AI, yet simultaneously predicts that 70% of emerging roles will be AI-related. Similarly, a South Korean study anticipates that 27% of jobs will experience hardship as a result of AI whilst 24% will experience higher productivity and wage growth.
We thus find Marx’s theories surrounding machinery accurate. AI is indeed out-competing workers in several fields, yet is also benefiting workers in others. In service and finance-centred economies, this development has been resoundingly negative, whilst those predicated on industrial capital have seen the inverse. Moreover, economies which are ‘split’ between these varying forms of capital, such as Malaysia and South Korea, are experiencing the full brunt of machinery’s boons and burdens. We therefore reach the following conclusion: that capitalism, not the technology it employs, must be the target of our critique. If Marxists adopt an anti-AI stance, we risk alienating the workers of sectors and economies currently experiencing growth from AI; though we also run the inverse risk by adopting a pro-AI stance. The correct approach is then to oppose capitalism itself, using the developments of AI to show its callousness toward those it no longer deems useful, as well as how its technologies can be appropriated to serve the proletariat’s historic task. To understand this latter point, we must also consider how machinery simplifies production. This readily applies to AI- it is doubtlessly simpler to prompt an AI to create a song than to compose one by hand. Marxists have long expounded on the concept of simplified production, as it plays a key role in the development of capitalism, the development of socialism, and the transition from socialism to communism.
Lenin recognised that capitalism tends to simplify both production and administration, which will be crucial in allowing the proletariat to manage the economy and state. He also noted that this is critical to the construction of communism. Only when production becomes so simplified that anyone can produce anything can the division of labour finally wither away. Only when the division of labour fades away can scarcity and a state which distributes products under such pretext vanish. Thus, he argued, the simplification of production is not only an element which empowers the workers to better manage the economy and state, but also one that will be key in the transition from socialism to communism. To be sure, this conclusion regarding the transition to communism was one originally drawn by Marx in 1875, though Lenin certainly expanded on the discussion.
Lenin was not the only theoretician to pursue this line of reasoning. Rosa Luxemburg approached the problem from a similar perspective, but outlined it in a manner which serves to guide our response to AI as it pertains to the petite-bourgeoisie. She wrote:
“Now, what is the economic significance of the extension of the system of shareholding societies? Economically, the spread of shareholding societies stands for the growing socialisation of production under the capitalist form – socialisation not only of large but also of middle-size and small production. The extension of shareholding does not, therefore, contradict Marxist theory but on the contrary, confirms it emphatically.
“What does the economic phenomenon of a shareholding society actually amount to? It represents, on the one hand, the unification of a number of small fortunes into a large capital of production. It stands, on the other hand, for the separation of production from capitalist ownership. That is, it denotes that a double victory being won over the capitalist mode of production – but still on a capitalist base.
“What is the meaning, therefore, of the statistics cited by Bernstein according to which an ever-greater number of shareholders participate in capitalist enterprises? These statistics go on to demonstrate precisely the following: at present a capitalist enterprise does not correspond, as before, to a single proprietor of capital but to a number of capitalists. Consequently, the economic notion of “capitalist” no longer signifies an isolated individual. The industrial capitalist of today is a collective person composed of hundreds and even of thousands of individuals. The category “capitalist” has itself become a social category. It has become “socialised” – within the framework of capitalist society.“
We thus see that innovations in capitalist production liquidate not only the workers of a given field, but capitalists, too. We can quite easily see how our aforementioned reading of Marx affirms such a view in the case of the petite-bourgeoisie. Luxemburg understood that innovations, such as shareholding, allow capitalists to centralise their fortunes whilst separating ownership from the day-to-day management of their capital. Such inventions, put simply, ‘socialises’ production by outsourcing the roles of capitalists to others- in shareholders’ case, this was to managers, foremen, and supervisors, in a word, workers, whilst in AI’s case, it outsources their role to machines. Moreover, it is clear, whether we use Marx or not, that machines out-compete human labour. We then find that AI can render many small producers superfluous.
How is this so? Many of the most emphatic opponents of AI are petite-bourgeois artists, entrepreneurs who sell commissions of their drawings, music, writings, etc. to fund their own existence. These individuals operate as small proprietors. They create intellectual property which they sell to others. The only difference between them and larger, more clearly-bourgeois artists, is scale, but both share the same relationship to production. Both create their own intellectual property, then either self-produce commodities relating to it or outsource production to companies which produce commodities, but either way, they self-produce capital and sell it to others. It is precisely these artists who deride AI the loudest, as it is their capital which stands to be rendered obsolete by the technology.
We cannot afford solidarity to this segment of the anti-AI left. Unlike the aforementioned proletarians, whose labour power is rendered useless by technology, these are capitalists whose capital is made useless by technology. Their criticisms of AI, though they may share the same features as proletarian criticisms, only arise out of a temporary overlap in circumstances. All the same, they stand by capitalism and the relations it promotes. They stand by intellectual property, as it is the basis of their right to capitalise on their artwork. A plurality of them tolerate the exploitation of workers as third-party firms produce their merchandise. They demonstrate their petite-bourgeois ideology for what it is as they criticise ‘capitalism,’ often condemning ideological trends, policy decisions, and egregious examples of exploitation, all whilst failing to recognise that their own creations are capital, that they perpetuate the alienation of labour, and that the only distinguishing factor between themselves and the capitalists they condemn is, once again, the scale of their operations. They are anti-capitalist capitalists, condemning the outgrowths of bourgeois society without ever assailing the mode of production itself- which they resolutely support.
They thus, alongside the less class-conscious workers, represent the reactionary segment of the anti-AI left. This segment, either owing to its lack of political education or petite-bourgeois status, opposes the development of technology outright. This is an incorrect stance to take. As Marx noted, it is not machinery, but the mode of production that weaponises it against the proletariat, that is the enemy of the workers. As Marx and Lenin noted, technological progress simplifies production, a reality which we can observe, and thereby makes easier the task of workers’ direction of the state and economy. Moreover, the simplification of production is a key element of the transition from socialism to communism. Luxemburg took this further, noting that certain innovations in capitalism lead to the superfluousness of the bourgeoisie. They increasingly develop into a class of alienated idlers as the proletariat carries out both production and the management thereof. Lenin also recognised this trend as it relates to imperialism and the increasing geographic alienation between production and the capitalists who ‘manage’ it.
Each of these concerns is relevant to the AI debate, illustrating both the damage being wrought by this technology on segments of the proletariat, as well as its capacity to serve the long-term interests of the working class. On the one hand, we must construct a programmatic response which serves workers who have been displaced by the technology. We must be able to answer their questions and serve their needs. Yet on the other, we must not stoop to trade unionist economism; we must not pit the workers of industries gutted by AI against those experiencing growth from it. Similarly, since the displacements and benefits of the AI industry are stretched across the globe, we must not reduce ourselves to nationalistic calls for autarky. The demonisation of ‘foreigners’ and their labour is, and always has been, a reactionary response to the contradictions of capitalism. We must instead find a way to build class consciousness, organisation, and political education among all industries and reject all attempts to divide the international proletariat. Further, it serves us well to unmask the opportunistic uses of anti-AI sentiments by the petite-bourgeoisie. It is the small proprietors who represent the least developed and most reactionary segment of the bourgeoisie, and though their concerns may appear in-line with our own, this surface-level connection is the alpha and omega of our shared interests.
This is Part 1 of 2. You can read Part 2 here.
[1] Labour power is the commodified form of labour. Thus, when I say machines require less labour power to furnish compared to humans, what I mean is that less labour needs to be purchased to support them. Machines, on the one hand, are commodities, labour power crystallised in a physical form. They can therefore be bought, and generally don’t require much to be spent thereafter. Workers, on the other hand, require not only the upfront investment of tools, workwear, safety equipment, etc. but also the recurring payment of wages, benefits, and so on.
PARTISAN! 
You must be logged in to post a comment.