A small community on the edge of Santiago, Chile, built a human-operated “chatbot” to dramatize the resource toll of artificial intelligence, fielding roughly 25,000 prompts over 12 hours and returning hand-drawn images and curated answers instead of automated outputs. The organizers in water-stressed Quilicura—now a hub for data centers from Google, Amazon and Microsoft—aimed to highlight AI’s often-invisible water and power demands, which vary by location and cooling technology. Google says its Quilicura facility is Latin America’s most energy-efficient and points to local water-restoration work, even as projects near Santiago have faced legal scrutiny over water use. The demonstration underscores the growing friction between AI’s rapid scale-up and climate and infrastructure constraints in drought-prone regions, foreshadowing higher compliance, mitigation and transparency pressures for cloud providers.
Related articles:
Making AI Less Thirsty: Uncovering and Addressing the Secret Water Footprint of AI Models
Microsoft will replenish more water than it consumes by 2030
