Destroy AI

Por Matthew Cole — Tribune

As tech companies accelerate their relentless drive to capture and control the flow of information across the globe, capitalism is entering a new phase. AI platforms like Palantir, Open AI and Anthropic are buying up swathes of land for data storage, deepening the automation of everyday life and the destruction of the climate. Unlike earlier waves of industrial innovation — laying railroads, building motorways, electrification — this new infrastructure brings no tangible benefit. Its costs far outweigh any potential gains. While AI executives have now begun to talk of ‘pacing the frontier’, the truth is that something much more radical is needed: to mobilise for the destruction of this infernal technology, and develop an emancipatory politics that rejects the datafication of the planet.

Against the Slopverse

AI is literally retarding our brains. If online media was once an echo chamber, it has now become a hellscape of algorithmic slop which heightens social alienation in ways that were previously unimaginable. A recent study shows that students who use LLMS to write essays exhibited reduced brain activity and diminished critical engagement compared to those who use search engines or simply their own unaided reasoning. Researchers at the MIT Media Lab found that the convenience of automated answers produces so-called ‘cognitive debt’: a clear, measurable contraction of human attention and creativity. Another study by the OECD’s Programme for International Student Assessment identified a general fall in numeracy and literacy comprehension that coincides with the mass adoption of these technologies.

Behind AI’s consumer-facing slopverse there is a material infrastructure of staggering scale. Nvidia’s global supply chain, for example, draws tungsten from Brazil, gold from Colombia, and tantalum from Kazakhstan to manufacture GPUs at TSMC in Taiwan. These chips travel through a planetary web of logistics operators, miners, data labellers and e-waste dismantlers. Over 10,000 data centres now operate across 170 countries, but their distribution is profoundly uneven. Because the United States, China and the EU host more than half, computational power is highly concentrated in the Global North. The geography of AI thus reproduces old imperial hierarchies. The Global South provides minerals, labour and land; the North monopolises compute, patents and profit. 

All this will get worse as the different Big Tech players struggle for dominance. Amazon plans to invest $150 billion on data centres over the next fifteen years, while Meta’s projected spend on data centres was $37 billion in a single year. The United Kingdom aims to build 100 data centres by 2030, including a £10bn AI data centre in Blyth consisting of 10 buildings covering 540,000 square metres. These will all consume vast quantities of energy and water, straining local ecosystems while producing tons of greenhouse gas emissions and e-waste. 

The global data cloud has a greater carbon footprint than the aviation industry. Energy consumed by data centres alone accounts for 0.3 percent of global carbon emissions. AI advocates argue that as more hyperscale centres get built, efficiency will improve and green energy will be used more — but similar ‘net zero’ pledges by European energy companies from Shell to BP have been exposed as simply crass greenwashing scams. There’s little reason to think AI will be any different. 

Digital Resistance

People across the globe are now recognising these social costs. Between 2023 and 2025, activists blocked $18 billion worth of data-centre projects and delayed another $46 billion across 24 US states. Outside the capitalist core, similar movements motivated by environmental and infrastructural concerns are taking off. In Cerrillos, Chile, communities led by local activist Pamela Ramírez halted Google’s water-intensive facility amid a historic drought. In Dublin, Ireland, the ‘Press Pause on Data Centres’ campaign has been fighting back against data infrastructure that threatens to consume 27 percent of national electricity by 2028, while in the Netherlands the Save the Wieringermeer campaign has defended farmland and aquifers from Microsoft projects. 

The resistance to data-centre development in the UK combines struggles over land, energy, water and urban space. From East London to Buckinghamshire and Fife, objections around housing, energy grid capacity, water, pollution and work converge as communities are expected to surrender socially necessary resources to privately owned tech infrastructure, to enable AI expansion that nobody wants. 

The dispute over Brick Lane in East London, recently written up in these pages, makes this especially stark. There, the government is determined to redirect resources that could be used for housing and other social needs towards a highly capital-intensive data centre. There are over 192 data centres operating or being built around London: infrastructure won’t provide jobs or benefits to the community, but will simply become another tributary to US tech oligarch fortunes. Any opposition to them is cast as an impediment to ‘growth’. While Brick Lane is being forced through by Burnham’s cabinet despite the local mayor blocking it, Buckinghamshire was successfully stopped.

Futures

In light of the dangers of Big Tech, the scholar Dan Mcquillan advocates what he calls a ‘decomputing society’, involving the dramatic reduction of datafication and computational energy from infrastructure to culture. Decomputing requires democratic limits on data extraction and energy-intensive computation, prioritising socially beneficial technologies over capitalist efficiency. The campaign known as ‘Welcome Spring, Burn a Tesla’ has targeted Elon Musk’s Gigafactory in Berlin along with his dealerships across Europe, and the Algorithmic Sabotage Research Group has drawn up a manifesto for algorithmic disobedience: disrupting AI systems through data poisoning, misinformation loops, and ‘tarpit’ traps that flood data-scrapers with nonsense content. 

These disruptive strategies are one thing, but building a compelling social alternative is another. The classic left critique of technology reminds us that it is never neutral: that relations of power are reflected in its design, and that it therefore reproduces the capitalist imperatives of competition, scale and profit. Today, with politics lurching even further right, new technologies are absorbing reactionary tendencies from society at large. As oligarchs advocate a return to monarchy and promote conspiracy theories about the Antichrist, their fantasies become embedded in the algorithms that they own. 

The history of class struggle from the Luddites to the Writers Guild of America shows that the emancipatory potential of technology never comes from the technology itself. It comes from building enough power to fundamentally remake these infrastructures. Hence the recent movement for digital sovereignty: public clouds and computational models with transparent safety frameworks which are democratically designed, rather than imposed by the superrich. The Non-Aligned Technologies Movement aims to translate this demand into an international campaign against the unequal digital world-system. 

For such movements to succeed, they would need to be aided by a state-led industrial strategy to decommission data infrastructure or reorient it towards socially necessary functions, set by workers and users. A process of collective decomputing could meanwhile ensure that all data contributions remain in open commons, controlled by everyone rather than enclosed by rentiers. This might seem like a distant prospect. But the fact is that any technology that emerges out of intellectual monopolism and extractive supply chains will be marked by these destructive forces. Remaking the former inevitably means destroying the latter.

Fonte: Tribune