

LSE IQ: is AI destroying the planet?
Aug 21, 2025
In this discussion, Anna Bevan, a producer of LSE IQ, reveals the hidden environmental costs of AI technology. Ergenie Dugois, an environmental economist, stresses the need for sustainable data center designs to curb energy waste. Sociologist Nick Caldry raises critical concerns about the social and environmental ramifications of AI, advocating for a sustainable social contract. Together, they explore the delicate balance between AI's potential benefits and its environmental footprint, highlighting urgent calls for policy and structural changes.
AI Snips
Chapters
Transcript
Episode notes
ChatGPT Water-Use Example
- Anna Bevan asked ChatGPT “how much water does ChatGPT use” and got an answer estimating ~500ml per interaction.
- She compared weekly use for a year to pouring away 27 litres, illustrating hidden water costs.
AI’s Computation Heats Infrastructure
- AI queries run huge parallel calculations which generate substantial heat and energy demand.
- That heat forces data centers to adopt intensive cooling, increasing water and power consumption.
Data Centers Can Devour Local Water
- Large data centers can use 1.5–2 million liters of fresh water daily, rivaling local human consumption.
- This creates strain on regional water supplies, especially where water is scarce.