AI Mysticism and the Climate Crisis
AI is damaging our planet far worse than we care to think or assume.
The use of artificial intelligence products, such as ChatGPT, Gemini, and DALL-E, has exploded in the past few years, and their environmental impact is becoming apparent. AI uses machine learning models that process massive amounts of data, and therefore require massive amounts of computing power to run. This makes AI prompts significantly more power-hungry than a traditional browser search. A prompt to ChatGPT uses 10 times as much power as a typical Google search. Multiplied across ChatGPT’s hundreds of millions of users, ChatGPT consumes as much power as a staggering 180,000 American households every day.
Using large amounts of power means that the computing equipment heats up and must be cooled down using water. Each conversation with ChatGPT consisting of 20 – 50 questions uses 500ml of water – as much as a standard water bottle. xAI, Elon Musk’s AI venture which created the chatbot Grok, is projected to use almost 4 million litres of water each day, or about 1.5 Olympic-sized swimming pools.
In addition to their environmental footprint, these data centres can have devastating effects on the communities they are situated in. xAI’s new data centre, located in south Memphis, Tennessee, installed 18 methane-gas generators without permits. Together, the generators can emit up to 130 tons of nitrogen oxides every year – a family of chemicals that can lead to respiratory diseases with exposure. In Granbury, Texas, the fans used to cool a data centre used for Bitcoin mining are generating so much noise that they are giving the residents of the neighbouring town sonic damage, with symptoms including loss of hearing, violent headaches, and nausea.
Yet, none of this is apparent when interacting with AI in our busy student and professional lives. Consumers of AI only see a clinically sleek interface, a bar where a prompt is entered and AI’s response to that query. The consumer is completely alienated from the physical processes of AI, and consequently, the trail of environmental and health damages it leaves in its wake.
The tech industry invites us to see its creations as immaterial. Putting your data on ‘the Cloud’ feels almost mystical, as if PDFs are being sent to some ethereal file void. Using AI chatbots seems more akin to consulting a friend than executing complex computer programs. This kind of language and attitude obscures the material basis for and environmental externalities of this technology. This perceived immateriality makes users less inclined to consider the very material consequences of using these so called “digital” platforms.
However, mysticism and suggested immateriality are not by-products of tech like AI but its foundation. To understand the development and creation of AI, it is crucial to situate it as a technology within the economic system that has incentivized and enabled its creation. Capitalism is an economic system based on infinite growth and can only exist where profits can perpetually be made and, therefore, capital can be accumulated. It is also a system that is in direct contradiction to the fact that Earth is a planet with finite resources — once the resources have been depleted there is nothing left to turn a profit off. This fundamental contradiction is becoming increasingly starker – and more dangerous — with the worsening of the climate crisis.
Technocapitalism, a term coined by political scientist Luis Suarez-Villa, attempts to grapple with this contradiction by commodifying intangible things like ideas, creativity, and research, instead. Under industrial capitalism, the resources exploited for production are overwhelmingly physical materials. However, the move to regard intangible resources in the same way considerably extends the list of exploitable resources, something that tech companies are more than eager to explore. These intangible things are difficult to quantify but are things that should not run out as long as humans are around.
AI seems to follow naturally from technocapitalism’s doctrine. With the rise of the internet, people are able to freely share their artistic creations and what’s on their mind. From illustrations to poetry, the internet has become an almost endless trove of creative resources that generative AI can tap into. With an enormous and free source of the intangible resources it requires (creativity) AI offers the prospect of endless and efficient production. Digital products can be created with what seems like significantly less human intervention and raw material input, marking a divergence from industrial production. The seemingly elusive nature of the material basis for AI is, to some extent, because some of the resources that it uses are indeed intangible.
However, it is a fallacy to believe that technocapitalism, and its products like AI, is truly distinct from the industrial and traditional means of capitalism can completely break with the industrial capitalist way of production. Just like industrial capitalism, the production of AI combines raw materials — such as fossil fuels and water — with human labour, such as programming, engineering, servicing data centres, and data labelling. Therefore, it is ultimately subject to the same constraints, and riddled with the same problems, such as the exploitation of natural resources and labour in the Global South. For example, in response to ChatGPT spitting out inappropriate responses, OpenAI outsourced its data labelling and moderation jobs to a firm in Kenya. Workers were made to read graphic and disturbing scenes described in detail for less than $2 USD per hour. These issues are not aberrations, but fundamental features of the production of these digital products.
Despite AI and the tech industry’s attempt to break from industrial capitalist-style production, it is ultimately unable to transcend the basic fact that products cannot be created without combining raw materials with human labour — that is, that something cannot be created out of nothing. We must remain critical of AI’s environmental and social implications and look through its façade of immateriality. Although AI has some exciting and promising applications in fields such as weather forecasting and biology, it is also just another product that plunges the Earth deeper into the climate crisis.
