Communia

What is Artificial Intelligence today?

What is called "Artificial Intelligence" actually has nothing to do with human intelligence. AI is a set of systems for statistical analysis of large masses of data that, based on models, adjustment, contrasting and a series of objectives, tries to predict a "correct" solution.

When an AI translates, for example, it is not translating in the sense that a translator would. It does not first understand what the text says and then try to convey the same message in another language by applying its own grammar and vocabulary. It simply tries to statistically predict, based on analyzing millions of translations, which sentence in language X would correspond to another sentence in language Y.

Like programs like ChatGTP, after raking through millions of conversations and interactions between users and web browsers, it predicts what someone asking a certain question or giving a specific general command expects - or at least can expect - to receive.

What has the "AI revolution" produced?

The procedures used by these systems have evolved over the last fifteen years thanks not so much to research, but above all to the reduction in the cost of processing speed and capacity thanks to the massive capitalization of large data centers. Without large computer farms, capitalized en masse by large online services we would not have neural networks like we have today.

That is, AI has developed as a product of the giant concentrations of capital that, in both the US and China, have shaped the Internet as a destination for large masses of capital in search of a destination.

It is no accident that the so-called Sputnik moment of AI that raised the competition between China and the US to the pre-war level we have now and pushed the US into a strategy whose first big pull was the war in Ukraine and the rising tension in Korea and Taiwan, started with an innocent Google-developed Go game program.

Nor is it any coincidence that the main US vanguard of AI service development, the OpenAI consortium, the creator of ChatGTP, was born from the initiative of Elon Musk and investors such as Peter Thiel with the support of Amazon and Microsoft. Behind the rapid development of AI are the large capital funds that once moved towards the Internet and online services with all their star companies and gurus.

What does AI really mean?

AI-controlled robot in an industrial process.

The development of AI is the second movement of what we saw with the Internet: the subordination of the general tendency to socialization of production to the placement and accumulation of capital.

Behind each AI there are many hours of hyper-precarious labor, but above all the corpus, i.e. the texts, conversations, behavioral data and choices of millions of people, i.e. many hours of social labor, including that which, not being paid, is not usually counted as such.

In other words, it is a forced and, as we shall see, violently contradictory attempt to confine the tendency and necessity to socialize production within the narrow margins of capitalist relations.

In reality, the emergence of the AI industry, as before it the emergence of large Internet monopolies, shows in an extreme form that this tendency to socialization is reaching its limit in capitalist society.

We are at a time when the very idea of machine is being transformed to openly become the materialization of social labor with no longer any justification for the need for wage labor, and in fact in comes into direct contradiction with it. The AIs, which are after all machines, computer networks...

They are organs of the human brain, created by the human hand; the power of knowledge, objectified. 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. To what degree the powers of social production have been produced, not only in the form of knowledge, but also as immediate organs of social practice, of the real life process

Karl Marx. Foundations of the Critique of Political Economy (Grundrisse), 1857-1858

What does this mean? That the type of socialization of production that allows the development of technologies such as AI is precisely that which makes it possible as an immediate objective to liberate society from wage labor, commodification and scarcity. And for the same reason, the type of contradictions that these technologies amplify in the system make it necessary and urgent for Humanity to overcome capitalism.

But let us go by parts...

Why is big capital betting so heavily on AI and how does it affect workers?

As we saw in the series In Communism..., the socialization of production is the historically necessary engine for multiplying the productivity of labor (and resources).

Capital realizes this and, within the narrow framework of accumulation, understands the opportunity and enjoys the immediate prospect of making profitable large infrastructures without a single worker or reducing to a minimum the labor force in sectors such as textiles.

But what is most stirring the imagination of the bourgeoisie is the de-qualification of labor.

In the industrial world they are thinking of reducing the training time of workers who would become mere assistants to the machine. And in the field of services in completing the proletarianization of sectors until not long ago professionals (=petty bourgeoisie) that had already been absorbed into the salaried corporate structure and gradually made precarious: doctors, [junior lawyers of large firms](https://www.economist. com/business/2018/07/12/law-firms-climb-aboard-the-ai-wagon), code programmers of little-value, writers of clinical trials and other corporate materials, telemarketers, investment advisors or designers who would move from increasingly standardized, low-input jobs to being directly replaced by generative AIs.

The growing contradictions of a capitalism with AI

AI-guided sentry intended for infantry, border and, eventually, riot combat.

The general framework: capitalism that reaches AI is already reactionary and cannot use it for the benefit of Humanity.

From the point of view of production, the application of AI means an increase in the physical productivity of labor: an almost general reduction in the hours of social labor required to produce a certain quantity of services.

But the bourgeoisie is not interested in physical productivity per se. What it is interested in is the productivity of labor in terms of profit: how much profit it can make from an average hour of social labor. That means not only producing more, but also selling more... despite the fact that wages - the effective demand base created by capital itself - after such a technological change make up a smaller percentage of the value of total production.

In the ascendant capitalism of the 19th century this contradiction was resolved by an expansion of the world market and of capitalism as a whole, leading to a reorganization and increase in global production, an increase in the consumption capacity of workers as well as a reduction in the average working day. Progress.

But in today's decadent capitalism, with the world market reaching into every last corner, the system has no way to expand what Marx called the "restricted bases" of wage demand it creates.

Firms, by incorporating technologies that increase their physical productivity, cannot take for granted that there is a market that will absorb more product. So, in the aggregate, they simply reduce hired labor to increase profits, without total output growing as fast as it could.

That's not to say that the drive to win new markets doesn't redouble. On the contrary. It does. But those markets are already also capitalist markets with a national capital that plays the role of owner and which has exactly the same drives and needs for mercantile expansion.

That is to say, we live in an imperialist era and therefore each step forward of accumulation translates into an increase of imperialist tensions between national capitals. Competition soon ceases to be a peaceful game of prices and offers and becomes a strategic game of trade wars, currency wars, sanctions... and military pressures, if not war.

To sum up: the great technological improvements today, instead of progress, that is to say greater integral human development of the producers themselves, end up producing unemployment and impoverishment of the workers on the one hand and warlike tensions that push towards more misery and death on the other.

The contradictions of AI as seen by the bourgeoisie

The ruling class is not really concerned about the fundamental contradiction between what capitalist growth (accumulation) demands and human development. Moreover, it believes itself capable of restricting the minimal outbreaks of abundance that may arise at the edges of the market grotesquely deforming if necessary its own forms of property.

What worries it is the form that contradiction takes in its own terms. That is, that the potential productivity increase of AI evidences in an excessively rapid or violent way the shortcomings of effective demand growth that the system itself produces.

The corporate bourgeoisie is happy to improve profit by reducing jobs thanks to the use of AI in its enterprises... but it is not blind and realizes that the generalization of AI use, by the same token, reduces demand... and is worried.

Hence, for example, the debates on the impact of AI have, at first, revived interest in taxing robots or establishing a "Universal Basic Income". Proposals that, after a moment of media glory, were once again forgotten... because they did not solve anything.

"Universal Basic Income" would not alleviate the effects of robotization on unemployment. Robotization, like any technological improvement in production, produces unemployment if there is not a sufficient market in which to place a greater mass of product. But, as we have seen, that is precisely the problem of capitalism today, and UBI, like any income redistributive system, cannot "create new markets," or even significantly increase demand.

Robotization tendencies will not only continue but, in a scenario of trade war, will become stronger than ever in order to increase export competitiveness, and will generate more unemployment than ever - because foreign markets will tend to shrink for domestic capital.

Universal Basic Income, 6/5/2018

In the end, its way out is always the same: to step on the accelerator of war by trying to curtail the technological development - and especially that of AI - of their imperialist rivals by all means. And if necessary, as we see in Ukraine, to convert these technologies into weapons of war. We see this more clearly the more powerful the imperialist power is. Not only in the USA, also in China, which in fact is at the forefront of AI applications for the military-space race.

One could hardly expect anything that would summarize more clearly the growing antagonism between capitalism and human life: the tool that could serve to create abundance, becomes a tool of mass destruction and slaughter.

What can be expected from the statistical synthesis of capitalist morality?

A persistent tendency towards discrimination and atomization

With their typical and far from innocent myopia, the media have been insisting on AI biases as if it were their main problem, especially in generative and chat-linked AIs. Clearly, if the mass of texts used by a conversational AI is ideologically skewed towards sexism or racism, the AI will calculate the correct reaction to an interaction with exactly the same bias. And ex post corrections - prohibiting the AI from making jokes about blacks or women, for example, as ChatGTP does - will only serve to show the extent to which the discrimination game the system produces is pervasive because it will not erase racist remarks about Andalusians or insulting jokes about males.

However, the underlying concern about these issues on the part of the ruling class is actually more serious. They are already calculating the reduction of the management apparatus and the costs of the policies of aid to the vulnerable that now serve as an excuse to undermine the universality of basic services. But in the US, where it has first been tested at the local level, the AI has exposed the extent to which such decisions reproduced the classism and racism fed by the system itself. That is why EU legislators, among others, issue warnings about the moral messages that AI can reproduce and amplify and the European Commission produces guidelines to exonerate its responsibility in the use of weaponry controlled by these systems.

Automated repression

This is no mere futuristic fantasy. Lethal, autonomous military and police devices, without human control, are the unstoppable vanguard of the light weapons industry. Fully automated and autonomous weapon systems were known to have been tested during the Syrian war and reportedly increased lethality by escalating already bloody fighting locally.

From robotic dogs on the US-Mexico border to robotic sentries in India to armed armored vehicles in Israel, "automatic defense" systems are receiving increasing investment from states that go beyond warfare and are moving ever closer to automating the repression of civilians.

Totalitarian control of individual behavior

Among other things because the Chinese social credit system, a direct product of the union of totalitarianism and AI, originally criticized by the entire American and European press, not only found markets in Central Asia and Africa... but competitors in what they call the West. For example, South Korea tested its capabilities during pandemic confinements and Qatar turned the World Cup into a real-time demonstration of extreme AI-guided totalitarian control... in competition with Emirates.

In the US, both Google and Amazon are taking these technologies a step further, by monitoring "unusual behavior" among workers in the workplace... as if the increasingly widespread AI monitoring of work protocols and procedures were not sufficiently controlling - and sometimes dysfunctional and even dangerous.

A degraded and superstitious "universal culture".

Image created with OpenAI with the instructions: Running on the edge of space, towards a planet, quiet, reaching the abyss, digital art.

We are starting to see it these months with the social impact of ChatGTP: the regurgitated media of the opinion industry is starting to be seen as not only an informative, but an indistinguishable educational source, and Google goes "code red" in order to transform the answers of its search engine. If we heed the company's messages, we will soon get to a point where instead of receiving a list of diverse results, we will get wording more or less equivalent to a Wikipedia article.

If the relationship with information and knowledge mediated by Google weakened resistance to the ideology, propaganda and intoxication generated by the system itself, to the point of weakening social resistance against war, what is coming threatens to promote a level of homogeneity and demand for faith unknown since medieval Christendom.

It is no wonder that China is playing at replacing capricious influencers with AI-powered digital replicas. If the automation of social control refers to the social automaton, the mechanical utopia of the bourgeoisie, ideological AIs are seen by the ruling class as the soul of the social machine which, well controlled, would keep the great human meat-grinder running smoothly without any opposition.

This characteristically bourgeois idea of the soul of the automaton is not marginal. This summer Google fired its head of Artificial Intelligence for claiming that the company's equivalent of ChatGTP, on which he was working, had developed the perception and sensitivity equivalent to a child. The media response was not at all derisive. Everyone took it for granted that, even if the engineer in question had been wrong, it was only a matter of time, not of the nature of what we call AI, that these programs would become intelligent beings.

We are not here only facing a superstitious confusion or a deification of the statistical vulgate of the media, gigantic version of Vox populi, vox dei. When their reference newspapers confuse a statistical prediction in the form of an image with Art, we have moved beyond a typical phenomenon of the decadence of the system. We are facing the first steps of a brutal degradation of the culture of the ruling class itself, which once made the emergence of a universal literature and music the ideological ensign of its capacity to generate progress.

What should we expect from AI?