Which class is artificial intelligence for?

We are talking about artificial intelligence (AI) a little more every day. Some speak of the greatest technological revolution facing humanity, others of the millions who will be left unemployed, and some of the possibility that machines might one day destroy humanity. In my opinion, another question stands before all these debates: In whose hands will AI be, and whose interests will it serve? Because no technology is independent of the social order in which it is born.

The steam engine accelerated the birth of industrial capitalism. The factory created a new class relationship. The internet did not just accelerate communication; it reshaped the economy, politics, and daily life. AI is no different. Moreover, we are on the threshold of an even greater transformation. Because AI is increasingly thinking, classifying, making decisions, producing information, and accessing the memory of humanity. This means that AI will turn into a matter of power rather than just a matter of technology.

The words of Jacob Coxon, who recently resigned from Anthropic, are important for this reason. As a researcher who conducted pre-training research at both OpenAI and Anthropic, he stated that the two companies are in a race toward self-improving superintelligence and that this race puts human life at risk. Even more striking was that Evan Hubinger, one of Anthropic's safety researchers, agreed with the basis of this concern.

Why is the fate of humanity being left to competition between companies?

In the digital technology market, one company does not want to fall behind another. Investors want to grow faster. States fear losing technological superiority. Even if everyone is aware of the risk, the race continues. Because the logic of competition in capitalism often gives birth to the thought, “If you don't do it first, someone else will.” In this context, when the future of AI and class struggle become inseparable, we face two different world possibilities. First, for example, AI in the world is concentrated in the hands of a few new techno-bourgeoisie. Books, archives, scientific studies, works of art, and the cultural memory of humanity turn into raw data. The data, algorithms, chips, and computing infrastructures become theirs. Humanity's thousands of years of scientific and cultural accumulation feed their systems. And in the end, humanity becomes dependent on the infrastructure of these companies to benefit from the knowledge it produced itself. This creates economic and cognitive inequality.

However, we cannot say that the interests of the new techno-bourgeoisie-oligarchy and state powers are exactly the same here. On the contrary, there can be serious conflicts between them. But centralized information infrastructures can provide both large companies with economic power and states with unprecedented surveillance capacity. One wants to increase profits. The other, control. But both can gain power from the centralization of information. The interest of humanity, however, is the exact opposite.

How can the collective memory of humanity turn into the private property of a few companies?

More importantly, when physical resources disappear, who will audit whether their digital counterparts are complete and unaltered? Having a book scanned is not the same as owning the book itself. Because culture does not consist only of words. The edition of the book, its history, footnotes, notes scribbled in the margins, changes between different editions, and its relationship with other books are also part of culture. While the possibility of accessing information appears to exist, the ability to independently verify information will actually decrease. Yet, truth stands on verifiable information.

But another world is also possible.

A world where AI is not the private property of a few companies, but a part of humanity's collective production capacity. A world where it is used to save people from mandatory and arduous labor. A world where, as productivity increases, the wealth of a few billionaires does not grow, but the society's working hours decrease. A world where information is more accessible, science is not hidden behind patent walls, closed-box technologies do not exist, and education, health, and culture are more accessible to everyone.

When the wealth produced by machines is shared with the whole society, AI does not have to be an enemy of humanity. On the contrary, it can turn into one of the greatest tools of liberation in human history. Beyond the economic necessities that force people to work, it can provide people with more time, more education, more creativity, and more freedom. However, it would be naive to think that this will happen on its own. Here, which class controls it will be decisive.

In this context, the class struggle will also determine the role of AI on the fate of humanity. I think the AI era will be shaped around this question. AI will not provide the answer. We and our struggle will.