Technology is often regarded as the primary driver of development in capitalist society. The very dynamics of capitalism compel the entrepreneur — the personification of capital — to constantly reconfigure working conditions in order to boost productivity.
However, technological revolutions do not merely impact the workplace or the factory floor; they reverberate throughout the entire social fabric, changing living conditions, consumption patterns, customs, infrastructure, institutions, and more.
The invention of semiconductors (silicon wafers capable of amplifying and switching electrical signals) and the miniaturisation of electronic circuits — accompanied by a parallel increase in computing power — lies at the heart of the digital revolution we are currently experiencing.
Since the 1980s, new information and communication technologies have rapidly succeeded one another and successfully established themselves in the marketplace: personal computers, the Internet, mobile phones, digital platforms, and social media have permeated the social fabric and every aspect of our lives — from work and communication to leisure. State-of-the-art digital systems, driven by the development of artificial intelligence (AI), are likely to have an even more extensive and profound impact on economic and social life (some even speak of an "ontological revolution" — a fundamental upheaval of the structures that constitute human reality).
The term "artificial intelligence" was first used in 1956 at a famous conference held at Dartmouth College in New Hampshire. The event was organised by four researchers — John McCarthy, Marvin Minsky, Nathaniel Rochester, and Claude Shannon — who would go on to play pivotal roles in the development of this emerging field. Artificial intelligence can be defined as the discipline that studies the techniques, methods, and tools used to simulate — rather than emulate — natural intelligence, specifically human intelligence. (Advances in neuroscience have, for the time being, shed only partial light on cognitive functions; thus, the term "simulate" is preferable to "emulate”, as the workings of natural intelligence cannot be faithfully reproduced).(1)
Two essentially complementary approaches — symbolic and sub-symbolic — have been employed in the development of intelligent machines (in the sense previously described); both paradigms are applied in practice and should not be viewed as mutually exclusive alternatives.
The symbolic approach is grounded in formal logic, which, throughout its evolution, has provided a notation for the principles and expressions of reasoning. This formal system has enabled a mathematical foundation for symbolic reasoning, allowing knowledge to be represented as a set of symbols that can be manipulated by algorithms (i.e. sequences of instructions defined to perform a specific task or solve a problem). An example of symbolic artificial intelligence — or rather AI based on rules and explicit, rigidly predetermined knowledge — is IBM’s Deep Blue system, which gained fame in the 1990s for defeating world chess champion Garry Kasparov.(2)
The sub-symbolic approach, on the other hand, relies on methods that allow rules to be derived and generalised directly from data. In this case, it is essential to have a data set large enough and sufficiently representative of a given problem to build a model that can adapt to previously unseen situations, demonstrating a certain degree of autonomy in the assigned task. The cognitive process underlying this approach is inference, meaning the system applies what has been learned from previous data to make predictions or make decisions in new contexts (conclusions are drawn from known information).
Large linguistic models, such as ChatGPT, are trained on enormous amounts of textual data, which allows the system to learn the complexity and nuances of natural language. (Training an AI model involves teaching a computer system to learn through examples, rather than providing a list of predefined rules to follow). The ChatGPT model has proven capable of passing the Turing test (i.e., generating texts indistinguishable from those produced by humans), answering complex queries, and generating creative content based on the regularities expressed by the data, without an intrinsic understanding of their meaning. Therefore, a sufficient amount of data (typically a huge amount) and an adequate computational system capable of processing it are essential.(3)
Thus, by means of artificial intelligence machines are being built that are capable of thinking without a brain (or, more precisely, simulating thought), where knowledge is expressed through algorithms.
== In the Realm of Technology
Military applications (highlighted by the conflicts in Ukraine and the Middle East), new forms of social control, the growing concentration of power in the hands of the tech giants (increasingly integrated into the digital-military apparatus of the state), and the potential repercussions on the labour force and the environment are all generating more than justified apprehension regarding AI.
Let us briefly consider an aspect that our political movement has always deemed central — without, of course, dismissing the others as negligible — namely, the repercussions AI will have on the world of work. It is central because discussing work in this context necessarily entails evaluating the relationship between the workforce and capital, as well as the impact these changes may have on the entire process of social reproduction.
The computerisation of production is certainly nothing new. For decades, we have witnessed a comprehensive restructuring of the economy driven by a radical shift away from organisational practices in the factory typical of the Fordist model.
The introduction of microelectronics allowed factories to abandon the rigid structures of the assembly line and organise production according to models better suited to fluctuations in the market demands. Responding to market demands in real time, minimising inventory and associated costs, diversifying products based on target markets: these were the most visible effects of introducing information technology into production.
The 1980s saw the rise of "Toyotism," a factory organisation model based on the concepts of "Just-in-Time" production and Total Quality Management.
Computerisation and robotisation made it possible to reduce the number of machine operators and yet gain significant increases in production output. At the same time, the number of personnel dedicated to programming and control increased.(4) Above all, microelectronics has enabled the relocation of entire production sectors to geographical areas with significantly lower labour costs than in advanced capitalist regions, paving the way for the widespread devaluation of labour that has accompanied the so-called era of globalisation.
AI represents a further advance — both quantitative and qualitative — in the automation process that began with the introduction of information technology into the manufacturing sector.
For some years now, we have been hearing about "dark factories" — fully automated facilities staffed exclusively by robots, where both human workers and lighting have become superfluous costs.
A prime example is the "smart factory" opened in 2024 in Changping (an administrative district north of Beijing) by the Chinese multinational Xiaomi, a consumer electronics company. This facility is capable of producing a smartphone every three seconds, around the clock, achieving a total annual output of 10 million devices with virtually no human workforce.(5) The factory is fully automated, with only a few technicians to monitor the system. Through advanced sensors and machine learning algorithms, the robots are able to gather environmental data, diagnose problems, optimise processes, and even self-optimise during production.
"Dark factories" are not limited to high-tech giants like Xiaomi or Foxconn; they are rapidly expanding across all industrial sectors. Examples include a massive factory in Xinjiang (a Chinese autonomous region) capable of operating 5,000 AI-driven looms,(6) as well as the Zeekr (a Chinese electric vehicle manufacturer) facilities in Ningbo and Quzhou.(7)
Since 2025, the Port of Shanghai has been transformed into a robotic city where containers — fresh off the ships — are scanned by digital systems, transported by autonomous vehicles to sorting areas, and subsequently placed onto trucks or trains, or into specialised warehouses, by specialised robots. Driverless trucks and trains navigate the docks of the massive port guided by AI systems; the coordinated movement of these vehicles, managed by digital algorithms, has boosted productivity by 30–50% while drastically reducing accidents.(8)
It is no coincidence that we began our investigation into the impact of AI-driven automation on manufacturing with China. The country's massive volume of public and private investment has enabled Beijing to secure a leading position in high-growth, high value-added technologies, including batteries, solar panels, wind turbines, electric vehicles, and advanced robotics. According to the International Federation of Robotics (IFR), the rise in industrial robots installed in China has been massive, increasing from 189,000 in 2014 to over 2 million in 2024.(9)
With 567 robots per 10,000 workers, China ranks — in terms of manufacturing automation — well above Germany (449), the USA (307), Italy (237), and the United Kingdom.(10)
Beijing’s central government is encouraging automation through local subsidies and incentives. The “Made in China” programme includes tax breaks that can cover up to one-fifth of the expenditure on industrial robots. China’s goal is to maintain a central position in the global manufacturing market and preserve production capacity in a rapidly ageing country.
Economic competition is inevitably accelerating restructuring processes in “Western” capitalist countries as well. Following the same development trajectory, Siemens has introduced advanced automation systems at its Amberg plant in Germany, Tesla has followed suit, and Amazon has increased the number of robots in its logistics centres from 350,000 in 2021 to over 750,000 in 2023.(11)
The Port of Shanghai thus represents the future of logistics, just as Xiaomi’s or Zeekr’s factories represent the future of industrial manufacturing.
The impact on employment is easily foreseeable: while there is a need for a small number of skilled technicians dedicated to remote supervision (capable of controlling multiple facilities from digital centres) and the maintenance of robotic systems (along with cybersecurity experts), the number of production workers is set to undergo a drastic — or rather, dramatic — reduction. In 2016, the Taiwanese electronics giant Foxconn replaced 60,000 workers with robots at a single plant in Kunshan (a county-level city in China’s Jiangsu province).(12)
According to World Bank data, the manufacturing sector employs over 100 million workers in China alone. A 2017 analysis by Oxford Economics predicted the loss of 12 million manufacturing jobs by 2030 due to automated robotic systems.(13)
AI applications extend beyond factories and logistics to the entire world of work, particularly the service sector (in "Western" countries, the service sector employs approximately three-quarters of the total workforce); 80% of companies already use AI systems in at least one of their operations, and over 90% plan to increase their investments in the sector.(14)
It is easier for AI systems to simulate conceptual, abstract thought than forms of thinking that appear simpler but require perceptual activity. This is because perceptual intelligence entails the ability to interpret and integrate sensory stimuli (auditory, tactile, visual, etc.) that are difficult to conceptualise and, consequently, difficult to encode. In this sense, it is "simpler" for an AI system to execute a design project, formulate diagnostic hypotheses, or defend a defendant in court than to navigate physical space with ease or manipulate objects with dexterity.(15) Certain manual jobs are unlikely to be replaced in the short term (cooks, retail staff, truck drivers, hairdressers, painters, cleaning staff), whereas the roles most vulnerable to replacement are found in the advanced service sector — including highly skilled positions (analysts, financial technicians and specialists, programmers, proofreaders, and application, database, or software developers).
According to a Rational FX report monitoring hiring and firing in the technology sector, more than 202,000 jobs were cut in 2025 due to corporate restructuring driven by the introduction of AI systems.
In recent months, all major tech giants have announced corporate reorganisation plans centred on the adoption of AI systems and workforce downsizing. HP aims to cut between 4,000 and 6,000 jobs by 2028, while Meta plans to cut approximately 3,600, Microsoft around 6,000, and Amazon 14,000.(16)
In the US, the unemployment rate among recent graduates rose by 5.8% in the first few months of 2025, with even higher rates in IT and finance — fields heavily exposed to automation. Since 2024, European technology companies have reduced hiring by 73.4%.(17)
According to Dario Amodei (CEO of Anthropic, a leading generative AI company), “AI could wipe out half of all junior-level service-sector jobs and drive overall unemployment up to 20% within five years.”(18) Meanwhile, a widely cited article by Frey and Osborne (2017), estimates that approximately 47% of jobs in the US are at risk of automation.(19)
In 2023, the OECD Employment Outlook estimated that about 28% of jobs in member countries are highly exposed to AI (and thus at risk of replacement).(20)
Various studies highlight that new AI-based digital systems will impact virtually all professions. Broadly speaking, low-skilled jobs are particularly vulnerable to automation and robotics, whereas high-skilled roles appear more exposed to AI technologies that heavily impact cognitive skills. Among mid-level occupations, the roles most at risk are concentrated in the administrative and sales sectors.
== The Two Faces of Capitalist Production
Various intellectuals, economists, and commentators tend to downplay the impact AI might have on the world of work. It is argued that the goal of AI is to complement human capabilities rather than replace the worker. Consequently, the emphasis is placed on the need to redefine and develop new professional skills that enable workers to interact effectively with machines.
Others highlight that AI itself possesses the creative potential to generate new business sectors that could require the employment of skilled personnel.
Furthermore, new digital systems would make it possible to address demographic decline and the consequent reduction in the professional workforce needed to meet business and societal demands. AI would also enable the creation of care and healthcare systems capable of managing — at a reasonable cost — issues associated with an aging population.
In general terms, it is worth noting that technological innovation has historically generated more wealth rather than diminishing it; the very restructuring of the labour market, while potentially causing initial difficulties, would allow society as a whole to engage in more creative and recreational activities.
These arguments are, of course, not without merit, yet they tend to overlook the contradictions inherent in the current economic and social system. Under capitalism, the production process is simultaneously a labour process and a process of valorisation.
The labour process — to use Marx’s own words from Volume I of _Capital_ — is "purposeful activity aimed at the production of use-values … an appropriation of what exists in nature for the requirements of man … the universal condition for the metabolic interaction between man and nature, the everlasting nature-imposed condition of human existence".
Thus the labour process in itself — that is, the process aimed at producing useful objects to satisfy needs — is not unique to capitalism but exists across all historical eras; technical progress enables the worker (L) to set in motion (n) an ever-growing mass of means of production (Mp). Historically, technical progress and the development of productive forces are expressed through the growth of Mp relative to L.
From this perspective, AI — like many technological advances that came before it —increases productivity and, consequently, wealth (defined as the quantity of use-values or goods produced per unit of time).
However, under capitalism, the production process is also a process of valorisation. The capitalist invests money, not to produce objects that satisfy their own needs, but to secure an economic return — to make a profit.
The valorisation of capital — that is, the production of surplus value — represents, to quote Marx again, "the determining purpose, the driving concern, and the final result of the capitalist production process". This process of valorisation distinguishes capitalism from previous modes of production and is, therefore, specific to the current era.
By briefly examining the valorisation process, we introduce certain concepts that would require far more space to fully explore but which we here must partially take for granted. The starting point — which plays a central role — is the labour theory of value, according to which the exchange of commodities is based on the quantity of labour required to produce them (the value of a commodity is thus determined by the socially necessary labour time needed for its production).
To survive, workers on the market sell the only commodity they possess: their own capacity to work, or labour power. In this context, to produce and reproduce their labour power, workers simply need to live (requiring, obviously, food, clothing, etc.). The value of labour power is therefore represented by the means necessary for the workers' maintenance and reproduction.
In full accordance with the labour theory of value, workers sell their capacity to work in exchange for a subsistence wage. However, when employed in the production process, labour power exhibits a characteristic that distinguishes it from all other commodities: the ability to produce goods whose value exceeds that of the labour power employed. The process of producing surplus value is, therefore, nothing other than the production of surplus labour — the appropriation of unpaid labour within the production process; surplus labour that manifests and materialises as surplus value. The source of value is therefore not constant capital (C — that is, the capital invested in purchasing means of production, raw materials, etc.) but variable capital (V — that is, the capital invested in wages). Thus the process of capital valorisation can only take place through the employment of labour power.
Constant capital — which plays a decisive, and today we might say predominant, role in the creation of wealth (expressed in terms of the quantity and quality of use-values) — does not generate surplus value; instead, it merely transfers to the commodities a portion of the value consumed during the production process. In other words, machinery (including AI systems), raw materials, and everything other than living labour (i.e. labour power) transfer to the produced commodity only that share of labour consumed during production.(21)
Modern bourgeois economic theory — which, for the sake of simplicity, we shall call neoclassical — completely rejects the labour theory of value and centres its economic analysis fundamentally on market dynamics — that is, it is essentially structured in terms of an exchange economy; a famous definition of political economy by L. Robbins succinctly expresses this viewpoint:
"Economics is the science which studies human behaviour as a relationship between ends and scarce means which have alternative uses." (An Essay on the Nature and Significance of Economic Science, 1935)
Even some theorists who began from positions rooted in Marxism now consider the labour theory of value to be obsolete in contemporary society. They argue that, in the course of technological development, workers' knowledge and skills have been entirely expropriated by capital and incorporated into constant capital. In their analysis, capital tends to valorise itself (this is the position, for instance, held by proponents of the “cognitive capitalism” theory).(22)
First, in our opinion, this analysis confuses the labour process with the valorisation process. Professional expertise (know-how) is undoubtedly an essential element in the labour process (i.e., in the production of specific use values through a particular labour activity — what Marx calls concrete labour) — but it does not contribute to valorisation.
The constitutive element of the value of a commodity is represented by abstract labour. Abstract labour, again to use Marx's words, is "undifferentiated, socially necessary, general labour, entirely indifferent towards any particular content, for which reason it receives an expression common to all commodities and only distinguishable in terms of quantity … its independent expression as money, the expression of the commodity as price.”
Later we will discuss how technology can influence the organic composition of capital, and therefore also abstract labour. Meanwhile, these theoretical approaches, by failing to consider the valorisation process or tending to consider it obsolete, inevitably propose solutions that essentially lie in the sphere of distribution within the capitalist system. (For the moment, we are not considering the positions of those who believe that the current production system can be progressively abolished through technological development alone). Universal income represents one such response to employment problems, in an era in which work tends to become superfluous (or the famous "work less, work for everybody" slogan).(23)
Our political current maintains that the labour theory of value characterises capitalism and will maintain its objective validity in the economic process as long as the current mode of production survives. The crises of the capitalist system are manifested in the market, in the sphere of circulation, but in fact they are generated by the irreconcilable contradictions of the production process, that is, the valorisation process.
== Productive Forces as an Expression of Capital
Let us very briefly consider how the valorisation process influences the labour process. From its very beginning (when work practices were still those of previous modes of production) the productivity of capital has always manifested itself by the fact that the worker is forced to provide surplus labour, that is, to produce beyond their immediate needs.
From its initial development, capitalist production is distinguished by the fact that the means of production, the material conditions of labour (raw materials, means of labour, and means of subsistence) appear not as subjected to the worker but the worker as subject to them.
Capital employs labour. Labour is aimed at the valorisation of capital; it is not a means for the simple production of goods.
The capitalist does not hold personal domination over the worker (as occurred in slavery and under serfdom), but exercises dominion over the entire labour process, that is, over objectified labour, the product, the conditions of labour, and thus over the worker himself.
Even in this initial form, Marx argues, we witness a reversal of the relationship, that is, a personification of the _res_ and a reification of the person.(24)
The forms of socially developed labour that have succeeded one another under capitalism — namely cooperation, manufacture (as a form of division of labour), and the factory (as a form of social labour based on machinery) — present themselves to the worker as forms of capital’s own development.
Even the machine — that is, science realised (in the sense that technical progress is possible only through knowledge and application of the laws of nature) — appears as capital in relation to the worker.
All applications of science, the products of labour, and the very organisation of the labour process are entirely independent of individual workers; they stand over and against labour power as expressions of capital’s power and appear solely as means for the exploitation of labour and the production of surplus value.
Thus, when science is applied to the capitalist production process, it is aimed at fostering and maximising the process of capital valorisation by increasing the rate of surplus value — i.e. the rate of exploitation of labour power. In directly productive sectors, technology — through robotics and automation — increases productivity, while in distributive sectors it enables the rationalisation of capital circulation and the redistribution of surplus value based on criteria of resource optimisation and cost reduction (e.g., the automation of banking, insurance, and commercial services). Furthermore, science allows for the extreme simplification of human labour tasks whichever sector it is applied to. Technology thus simplifies and redistributes the most elementary functions in the workforce, resulting in an overall de-skilling.(25)
With the introduction of AI, the process of reification and de-professionalisation — which until a few years ago seemed confined to manual labourers and, to some extent, white-collar workers — is now affecting, and will increasingly affect, broad swathes of intellectual professions. AI systems can be viewed as forms of reified knowledge — that is, knowledge separated from intellectual labour, introduced from the outside, and crystallised into a material apparatus embodied within a machine.
The modern intellectual worker is thus confronted with pre-existing knowledge, expressed through algorithms; they can neither alter nor intervene in this knowledge, yet they must be able to endlessly reproduce it.(26)
The scaling back of university curricula represents an attempt to align education with labour market demands. Thus an overall reduction in knowledge and a restructuring of skill sets are aimed at enabling interaction with — and adherence to the directives of — AI systems (which have now become the true repositories of knowledge).
While technological advances naturally raise the level of expertise required to truly master these systems, such training is increasingly concentrated in a few centres of excellence or else managed directly by technology companies themselves.
AI systems are consequently shrinking the workforce across nearly all economic sectors whilst simultaneously de-skilling — and tending to proletarianise — the world of work.
== Productivity and Value Creation
Capitalists are constantly driven to reconfigure methods of production: cutting costs allows them to compete effectively and capture new market share.
By boosting productivity, a factory can manufacture a specific item using, on average, less labour than its competitors; the individual value of its product will therefore fall below the market average. This difference in value, which affords a competitive advantage, is the incentive that constantly drives the capitalist to increase labour productivity and, consequently, to continually revolutionise the very organisation of the labour process.
In the most productive company, labour operates as enhanced, exceptional work, thus delimiting the necessary labour time to the benefit of surplus value.
The resulting increase in productivity therefore acts as a generator of relative surplus value even when the use value of the goods produced does not fall within the scope of the workers' own means of subsistence. (Although technological innovation can determine the de-skilling of the workforce, it tends to increase the surplus value that is generated by the individual worker, which, in turn, increases exploitation).
The growing productivity of labour is expressed, as we have seen, in an increasing mass of means of production that the individual worker manages to put into operation in the work process. This aspect is expressed by the technical composition of the working process, i.e. through the relationship between the growth in the mass of means of production and the mass of labour power. This increase in the technical composition is reflected, once again using Marx's words, in “a higher [organic] composition of capital, i.e., the relative decrease of the ratio of variable to constant capital.” Thus, while increasing the relative surplus value, the historical process of development of the productive forces, stimulated by the same dynamics of valorisation of capital, ends up drying up the only productive source of value and surplus value, i.e. the variable part of capital (v): the workforce.
The tendency for the rate of profit to fall (i.e. the ratio between the extorted surplus value and the mass of total capital invested, in its different components) expresses the growing difficulty of valorisation given the development of the productive forces of capital, based on the increase in relative surplus value.
In the long term, the rate of surplus value (i.e. the increase in surplus value generated by higher productivity) cannot compensate for the effect that the change in the organic composition has on the rate of profit. This is because, Marx points out, the organic composition can increase infinitely while the mass of surplus value finds its limits in the duration of the working day.
Artificial intelligence undoubtedly represents an important factor in the development of productive forces (just think of the degree of automation it is capable of generating in the industrial sector) but, for this very reason, it constitutes a formidable obstacle to the process of capital valorisation, and therefore to the reproduction of the social system. Automated companies, precisely because they are practically devoid of labour power, undoubtedly generate wealth through the production of goods but do not generate surplus value. (In this case, the surplus value is drained from the sectors that employ variable capital — we must remember that the individual capitalist never receives the surplus value produced by his company but a share of the entire social surplus value produced by the overall exploitation of the workforce).
Marxist analysis therefore highlights the contradictory character of capitalism, that is, the conflict that arises between the material development of production and its social form. Only by freeing ourselves from the capitalist mode of production and the need for valorisation that characterises it, will we finally be able to ensure that the achievements of science and technology are used to respond to the needs and development of the entire society.
-- GS _
Battaglia Comunista _
June 2026
== Notes:
(1) Gennaro Vessio, _Intelligenza artificiale per curiosi_ (Artificial Intelligence for the Curious), p.14, Edizioni Dedalo
(2) Ibid. p.19
(3) Ibid. p.22
(4) Mauro Stefanini Jr., "Dopo la ristrutturazione la nuova composizione di classe. Verso la ripresa delle lotte proletarie," _Prometeo_, 6 December 1993, English translation: "Where is the working class? An analysis of class composition in the course of the crisis in Italy" in _Internationalist Communist Review_, no 12, 1994, https://www.leftcom.org/files/internationalist-communist-review-12.pdf
(5) Gianluca Riccio, "Dark factories: le fabbriche fantasma producono già da sole, senza luci né operai" (Dark factories—ghost factories—are already producing autonomously, without lights or workers), https://www.futuroprossimo.it/2025/08/dark-factories-le-fabbriche-fanta…
(6) Robot King, "Cina: la fabbrica fantasma con 5000 telai e zero operai" (China: The ghost factory with 5,000 looms and zero workers), https://robohorizon.eu/it/news/2026/01/cina-dark-factory-automazione/
(7) Mario Piccirillo, “Un esercito di robot alimentati dall’AI: ecco come la Cina sta vincendo la guerra dei dazi” (An Army of Robots powered by AI :here is how China is Winning the Tariff War’), https://www.dire.it/23-04-2025/1143084-un-esercito-industriale-di-robot…
(8) HPadmid, "Shanghai: il porto che si guida da solo" ("Shanghai: the self-driving port"), https://www.hackerpunk.it/blog/shanghai-il-porto-che-si-guida-da-solo/
(9) Maurizio Carmignani, "Robotica industriale, la Cina fa paura: l’Occidente è in ritardo" (Industrial robotics, China is a formidable force: the West is lagging behind), https://www.agendadigitale.eu/industry-4-0/robotica-industriale-la-cina…
(10) Ibid.
(11) Riccio, op. cit.
(12) Ibid.
(13) Massimo, "L’era delle fabbriche senza luci: come la Cina sta silenziosamente riscrivendo le regole dell’industria globale" (The era of lights-out factories: how China is quietly rewriting the rules of global industry), https://www.altogain.it/lera-delle-fabbriche-senza-luci-come-la-cina-st…
(14) Ludovica Favarotto and Claudia Schettini, "IA-lavoro: l’equilibrio fragile del progresso" (AI and Work: The Fragile Balance of Progress), https://www.ispionline.it/it/pubblicazione/ia-lavoro-lequilibrio-fragil…
(15) Vessio, op. cit., p.23
(16) Favarotto and Schettini, op. cit.
(17) Andrea Daniele Signorelli, "Le intelligenze artificiali hanno iniziato a rubarci il lavoro." (Artificial intelligence has started stealing our jobs), https://www.domusweb.it/it/notizie/2025/06/23/intelligenza-artificiale-…
(18) Ibid.
(19) Randstad, “L’impatto dell’intelligenza artificiale sui lavoratori italiani”, (The impact of artificial intelligence on Italian workers), https://www.randstad.it/fondazione-randstad-ai-humanities/scarica-impat…
(20) Favarotto and Schettini, op. cit.
(21) On this topic, see also: CP, "Post-Capitalism via the Internet (According to Paul Mason) – Dream or Reality?", https://www.leftcom.org/en/articles/2016-02-21/post-capitalism-via-the-…
(22) Andrea Fumagalli, _Valore, Moneta, tecnologia_ (Value, Money, Technology), p.57, Derive Approdi.
(23) For more on this idea, see our article on universal basic income: https://www.leftcom.org/en/articles/2018-04-14/universal-basic-income
(24) Marx, _A History Of Economic Theories_, *Vol. 1, p.341, Newton Compton Editori.
(25) On this topic: Mauro Stefanini Jr., “General Trends in Class Composition", _Prometeo_, Series IV, September 1984.
(26) Enzo Modugno, _Il cybercapitale_, p.11, Manifesto Libri (we do not share many of the political views expressed by the author, but we are substantially in agreement regarding the reification of the intellectual worker's skills induced by AI).