Techno-solutionism: AI researcher Paul Schütze: “Technology is inherently social”

Mr. Schütze, you distinguish between two perspectives on artificial intelligence (AI). What are they?

On the one hand, there is the increasingly common perception of artificial intelligence as an algorithm-controlled technology that is used in the broadest sense as a useful tool for various purposes. These are specific applications, such as the most well-known chatbot ChatGPT, which can generate multimodal outputs based on texts, images or films that you enter into it. This is the understanding of AI as it is most widely received in the media and as it is taken up in most scientific and technical discourses. Another, perhaps more important perspective on AI is that technology is inherently social. Every technology is always socio-culturally embedded in our society. AI is not excluded. This means that we actually have to understand AI not just as a tool or as a technological instrument, but as a system that is connected to large parts of society.

interview

Paul Schütze

Paul Schütze is a research associate in the Ethics & Critical Theories of Artificial Intelligence research group at the University of Osnabrück. In his work he deals with the social impacts and ethical challenges of digital technologies and the climate crisis. He has a background in social philosophy, critical theory, affect studies and recent perspectives in political economy.

You mention in one of your texts that AI could, with Karl Marx, be described as a general condition of production. What do you mean with that?

Some debates on the political economy of AI highlight that thanks to their versatility, similar to, for example, electricity, AI technologies are now applied or may be applied in the near future in almost all branches of capitalist value creation. This ranges from the automation of production processes in factories or the optimization of infrastructure such as power grids to the control of self-driving cars and their application for services. Now, however, I’m a little more cautious here. Especially in critical debates, one should not follow the current hype. Calling AI a general production condition runs the risk of understanding AI too much as an all-encompassing and transformative technology that we as humans are powerless to face. I find that problematic. Rather, the focus should be on the realization that AI should always be understood as a social phenomenon. And it is precisely this social interconnectedness that needs to be critically reflected.

What do you think of the term? »AI capitalism«?

I think one should understand AI as a result of capitalist production rather than seeing it as a new mode of capitalist development. The author Kate Crawford, for example, describes AI as a mega-machine and thus refers to the socio-material basis of AI, according to which AI is based on a whole set of technical conditions. These range from industrial infrastructure, supply chains and data centers connected globally via submarine cables, to human expertise and physical work. This mega-machine is in turn embedded in the capitalist system or its result. The term AI capitalism should therefore not describe a fundamental change in the capitalist system, but rather a change of perspective with regard to the effectiveness and importance of AI within capitalism. It is not the case that AI capitalism has replaced fossil capitalism, for example, and represents a new stage of capitalist development. AI capitalism is more of a form of capitalism that exists parallel to other forms. However, it may require a different perspective, as in my opinion it further intensifies central capitalist dynamics.

Investments by large tech companies such as Google, Amazon, Meta and others in the development of AI systems have also increased. What do you hope to achieve from AI as a technology?

This is of course a difficult question that cannot be answered monocausally. But one dimension that particularly interests me is the fact that these companies not only produce technology, but also ideologies. In capitalism there have always been trends that equated social progress with technological progress. These currents, which promise to solve social problems through technological progress, can be roughly summarized under the term techno-solutionism. The idea of ​​a depoliticized technology, which is often propagated in these circles, is of course not tenable. Technology is always linked to ideas about what kind of society is considered desirable and what ideas for the future there are. And if you watch, for example, advertising videos from Google about their AI applications, you see that the ideas that these big tech companies have about human progress are very one-dimensional. The main thing here is to optimize all areas of life through ever better functioning AI, without questioning the basic prerequisites of capitalism.

In your work you also focus on the connection between AI and the climate crisis. What does this relationship consist of?

The first thing that can be said here is that there are very concrete hopes attached to AI as a technical tool for overcoming the climate crisis. For example, AI is intended to help make city infrastructure more efficient or improve weather forecasts in order to be able to predict natural disasters more accurately. These are of course good applications in themselves.

nd.DieWoche – our weekly newsletter

With our weekly newsletter nd.DieWoche look at the most important topics of the week and read them Highlights our Saturday edition on Friday. Get your free subscription here.

But?

The ideology of the techno-solutionists is reflected in this naive hope for AI. If you become aware of the socio-material connections between AI and capitalism, you realize that the systemic problems of the climate crisis do not even come into focus with AI solutions. On the one hand, the hopes for a better future that are based on AI fail to recognize the extractivist and colonial background of this technology, the development of which is based massively on the exploitation of resources. On the other hand, this idea often results in a naive idea of ​​a green capitalism in which profits continue to be maximized and resources exploited, but “sustainably” with the help of AI. Another point becomes clear here: the shifting of discursive and perceptible boundaries through AI. For example, AI is used to mine resources that are already running low, such as some rare earths, even more efficiently and in smaller pieces, instead of preserving them. The discourse around resource exploitation is diluted and the boundaries are pushed back. The actually necessary systemic questions about how a fair and climate-resilient society could be built without the exploitation of people and nature are pushed into the background.