Generative Artificial Intelligence has lent itself to be a technology of kinds immeasurably harmful to the ethical and productive considerations to the labour movement. Reactionary Marxists may set forth their arguments against such an idea, but we must be clear on why the narrative for G.AI is fundamentally a distortion of principles from the side of Marxism.

I will first get into the details, and later brief up the position against AI.

Firstly, we must define AI. What is it? Well, bourgeois interests define it in simplest terms, AI is artificial intelligence, cultivated by machines or algorithms that consume knowledge given to them and churn out solutions and answers to tasks and knowledge as it sees fit. This seems simple enough, yes? How different is it, then, to something like a printing press? or a machine in a factory?

To answer this as Marxists, let us understand better how these algorithms work, and how they are used in production. There are two different sides we must look at AI, both contradictory to one another (for it is dialectical, you see!)

First is its position by its material self — a plaything brought up by the capitalists to maximize production as well as for wholesale cutting of costs by the removal or gutting of labour from the mode of production. But then, how do you differentiate from the labour absorbed by AI and the labour absorbed by the usage of machinery and automation? ? One of which we must find acceptable and the other which we must not? Is this simply a vague abstraction from the view of a Marxist?

No. This position can be easily explained by how we can perceive AI by its use case, and can be further interrogated when we see its effectiveness in action. Let us remind ourselves of how said bourgeois interests define AI; which in itself is a melting pot of various different ideas, some of which are socially productive, some of which prove not.

Thankfully, this abstraction has already been solved for us, as the popular mind can differentiate AI as two different things: one of which is Generative AI, and the other is so called Machine Learning. For the sake of the text, let us refer to them as GAI and ML respectively.

Okay, we have now split AI down to its two main functions. For the sake of being thorough while beating the horse dead, let us define them as well.

Generative AI is artificial intelligence that takes the input given to it by an individual or group, and digests this input and stores it for future needs. When asked to solve a task or problem, it picks at all the inputs it has been given before and patches an answer together like a taxidermy based on its contexts, the inputs it has been given, and the particularness of the task at hand, i.e: to what detail it must truncate or bloat itself to.

Machine Learning is artificial intelligence that takes the input given to it by an individual, while containing within itself certain procedures and qualifications it must adhere to, after which it stores said input. After this, when a task is given to it, it follows its own internal logic and then provides an answer, solution or in some cases, information based on the prior input.

“Then pray tell, what is the difference between these two, since the definitions sound so eerily similar?”, One might ask. It is rather quite simple; GAI creates an output based on intangible, irregular and uncontrollable datasets, whereas ML adheres to simple rules that is known to all, is primarily sharpened for only one task — one that does not involve collaging together fundamentally different data in order to come to an answer.

One is a sharpened rapier with which one may pierce their opponent with great accuracy, made and refined for this task alone, while the other is a brutish, heavy, blunt sword which is made by the melting of various different swords each of which were perfectly fine on their own. Both swords need a person to wield them, in order to be effective. A hammer needs to be in the hands of a worker in order to make any good sword.

Now that we have taken this sidewalk away to define these tools, let us walk back to a position where we can analyse them clearly.
Returning to its use in production, we see GAI and ML playing two very different roles in the mode of production and under capital. Before we return to the first half of AI, let me get to the second half, for it is needed to better understand the prior.

AI, both GAI and ML under capital, are oppressive tools. They use both the value found in their cost-cutting nature to replace labour, and for undermining proletarian labour-power. So for this and perhaps when the discussion is about this, we must oppose them being used as such; tools where they discard the worker's labor without thought and action! We oppose the wholesale use of Artificial Intelligence as a whole, not because it is not the newer, pleasant progression in productive forces, and such would remove in part the need for labour, but because they are and will continue to be used for oppression in the field of capital and crude economics! As long as the workers do not own the means of production, Artificial Intelligence, whether it is GAI or ML, is simply just the next machines, the next trains, the next spreadsheet unless it is held hand-to-hand with the worker.

However, we must understand where AI has helped us - not in its material self, as many of us seem to think that asking AI for solutions to problems might seem more effective to us than anything else (refer back to how we had defined GAI, one that cobbles together answers from the input given to it, and only serves to give us a Frankenstein!), but rather its shift in relations between the worker and capital.

We are still on the eve of this development of technology. Perhaps the productive forces behind the development of AI cut it short in fear of it accelerating the worsening conditions of capital - Perhaps it will accelerate the contradictions inherent to capital such that capital itself will topple against the pressure - we cannot predict effectively how such a technology will develop. However, in the few years it has been deployed in production, we can understand that it has, and will have consequences on the relations of capital and labour.

AI has degraded the art and text written by its GAI counterpart with it. Like Machines, the value of art in all fields have in fact, degraded. This has usurped the roles of art and text in history - left them accessible to the masses, even at the cost of their value. In another text perhaps, this must be analyzed and talked about, whether the new roles of art and text under AI are productive or destructive to the health of a general, conscious mind. This in itself, within the few years of development that we have seen, has changed the relationship of the state, the capitalist, and the worker and the peasant.

We see in America, what can be called as perhaps the most visible abstraction and contradiction, the first we see in AI; where the AI-generated politicized material that is released by the state, whether for aesthetic purposes or not, has continually de-politicized art in its present form. No longer do the masses find artwork as a genuine platform for resistance. The American masses do not find themselves able to find themselves conscious enough to find meaning in politicized art. This is another point to be understood clearly; The use of AI as a weapon of the state against the masses in attempts to both de-politicize their inherent political existence, and to delay or outright work against their organizing.

We also see this in other countries where there is overt capitalistic attitude, such as India, where the state regularly organizes propaganda that is generated through GAI, depriving the homegrown art of its political value and directly influencing trends that seek to remove the art from its political roots. This is not a new trick for the BJP, it continues to depoliticize almost every political aspect of Indian cultural icons in order to suit their narrative.

This de-politicizing trend in art, which hasn't yet affected written text quite as hard, has made effective organizing incredibly hard - namely in the fact that no longer is art; in form of poster, in form of pamphlets or print, seen with the political vigour it had been seen decades or even just years ago. All trends in art fields have been moving towards of a point of dilution against any hint of political thought, or rather any hint of rational thought.

Let us revise the two ways AI is being used as a tool under capitalism then:
1. As a tool of oppression by cutting down wages for the workers as well as the dismantling of labor power, further alienating the worker from the surplus value.
2. As a tool for propaganda and the depolitization of art, text and all other various fields of communication.

With this, let us revert back to our first half of functions of AI; namely of its two halves, and brief up our positions.

When we do this, we understand two things;

The first usage of AI as a tool - as a tool of oppression only exists under capitalism. It is not oppressive when this is used as a progressive productive force by the hands of the worker. Another thing to note is that this role is played almost exclusively by ML applications and not GAI.

The second usage of AI as a tool - as a tool of depolitization of the daily life of the worker is harmful no matter what. It has been scientifically ascertained exposure to GAI content has long term damage to thinking faculties, something that we must not aim for if we are to have enlightened, class-conscious workers amongst our organizations. This role is played exclusively by GAI and not ML.

With this, it is understood which sides of AI we are to support the development of, and which to tear down, as well as understood precisely why we must do each of these.

There is of course more things to think about with AI, we are but on the brink of a revolution in production, and we do not know just how revolutionary it might be. Until we know of more developments, I believe with the strongest conviction this is the position we must take against revisionism in this topic.