Artificial intelligence and war: A fight for control

This Monday we start a new program on LID+ with “Chipi” Castillo and Jesica Calcagno. In this column we will try to open some discussions on that terrain that intersects the so-called new technologies and the digital world with politics and class struggle.

In this case we begin with a debate that is centuries old, but that since the beginning of the 20th century has become increasingly important: the relationship between technology and war. Many of the technologies we use every day, such as radio, Internet, GPS or aviation (just to name a few), were born or made great leaps driven by military development and the resources that States invested when it came to projects that were not profitable for private activity.

Today we are seeing something similar with artificial intelligence, which although it is not born as a state project, is getting fully involved in the debate about war and its use by States.

Obviously, many of these processes have their epicenter in the United States, which as the main imperialist country condenses within itself many of the central contradictions between the development of technology and the use that this capitalist society gives it.

Donlad Trump Vs. Anthorpic

Anthropic is Claude’s driving forcethe family of generative artificial intelligence language models that has gained a lot of visibility in recent months. Dario and Daniela Amodei They founded it after resigning from OpenAI, which is the driving force behind ChatGPTafter raising several differences about the security and approach that this company was taking.

In 2024, Anthropic signed an agreement with the Pentagon for 200 million dollars to integrate its artificial intelligence into its systems. The objective was clear: intelligence analysis, massive data processing and identification of military targets.

But the contract had two limits: the system would not be used for internal surveillance of the population and not for the use of completely autonomous weapons.

The problem began when the Pentagon wanted to eliminate those restrictions. He wanted to be able to use the technology for “any legal purpose” and Anthropic refused.

Donald Trump’s government responded by escalating the conflict: it canceled the contract and declared the company a “risk to the national security supply chain.” A category historically used against foreign companies accused of espionage, as has been done with the Chinese company Huawei. The novelty was its application in an American company.

What happened?

OpenAI quickly rushed to the doors of the Pentagon to take its place and the succulent contracts, and Anthropic has already appeared in court appealing the government’s action.

At first glance, it might seem like an ethical dispute. A “responsible” company facing a “limitless” State. But that is, at best, incomplete.

The company may talk about ethics, but it is also defending its business and its power. User growth, brand positioning, the fight for the market: all of that is at stake.

Claude herself responds that, despite the loss of the contract, the company is experiencing a “rebound effect”, since “daily active users have grown more than 140% since January.” In a polarized US, confronting Trump can also be a business. A kind of “human washing” where ethics also works as a business strategy.

They defend their power. Because these same technologies are already being used in real military contexts, processing intelligence information and participating in scenarios such as Venezuela or Iran.

When we said that they did not accept that Claude be used for internal investigations, let us understand that “internal” is for US citizens; The other inhabitants of this planet… I owe you.

And when they say that “they will not use our AI for autonomous weapons,” it is because “Claude is not ready for that yet.”

Open some background questions

Artificial intelligence works on data. This data is produced by all of us, all the time: when we use the phone, when we move around the city, when we buy something, when we watch a series, even when we use artificial intelligence itself.

The knowledge that allows them to be processed is based on mathematics, cybernetics, computer science, largely developments produced over decades in public and/or publicly funded universities.

But all of this ends up concentrated in a funnel controlled by private companies and some imperialist states.

There we have one of the core of the problem.

Because that concentration allows something that didn’t seem like science fiction so long ago: social surveillance systems capable of building detailed profiles of individuals and entire populations.

In the military field, systems that allow data to be crossed, detect patterns, identify objectives and accelerate war decisions.

Faced with this, the debate often slides towards a false problem: whether technology is good or bad. Technology is not “neutral” but it is not the one that decides either. Focusing only on it takes political decisions out of the scene.

For example, Israel has invested billions of dollars in artificial intelligence.

But in Gaza They decided on mass destruction and genocide. It was a political decision.

Now a lot of research is also coming to light that reflects the sexist and racial bias of AI responses. Whether by reproducing the same biases of a capitalist, racist and patriarchal society, or by editorial lines, it marks other problems to be addressed.

The problem is not artificial intelligence itself, although we must study and rethink it, but the power relations in which it develops.

It is a new layer in the dispute for power in capitalism. One where information, algorithms and the ability to process them become central tools to govern, monitor… and wage war.

And there the questions return, more and more urgent:

Who is going to control all this data that makes up our lives?

Private companies governed only by the thirst for profit?

The imperialist states getting ready for war?