To change the conversation, we must stop treating AI as a theoretical ghost and start treating it as a physical machine. Right now, most debates focus on whether AI will one day wipe out humanity or crash the stock market. This centers the debate on the survival of Western power structures. We need to pivot toward the immediate, material realities of how these models run.
Start by highlighting the extraction economy. This means talking about the lithium mines in South America and the massive water usage required to cool data centers in drought-prone regions. Instead of debating "superintelligence," we should discuss the soil degradation and local water scarcity caused by hardware manufacturing and energy consumption.
We also must address the invisible labor force. Large language models depend on thousands of low-paid workers in countries like Kenya or the Philippines to label data and filter toxic content. These workers face psychological trauma without adequate support. If we want a meaningful slowdown debate, we must demand transparency regarding energy footprints and labor conditions. We shift the focus from
