As the world giddily embraces artificial intelligence, a new United Nations report has thrown a bucket of tepid data centre runoff all over the notion that smart machines will eventually consume less energy. Spoiler: the machines are not coming to save the planet – they’re coming to rewire it, one megawatt at a time.
Jevons Paradox: Unwelcome Return
Tech enthusiasts might want to consult their Victorian history before promising salvation via efficiency. The report warns that ramping up AI’s productivity only lures society into using ever more of it, bringing the nineteenth-century “Jevons paradox” roaring back from obscurity. Instead of scientific progress yielding eco-dividends, every gain becomes an excuse to just plug in more of the stuff – more text, more images, more code, more thirst for electricity and water.
The vision of a fine-tuned, efficient AI future promptly collapses into an energy-guzzling feedback loop only a mathematician’s mother could love.
At present, planetary resource consumption is hardly a modest affair. In 2023 alone, data centres soaked up more power than entire oil-rich nations and slaked their vast mechanical throats with water equal to humanity’s annual drinking needs. Sounds like progress: for computers, perhaps, if not Homo sapiens.
The UN’s estimate that AI could soon consume 3% of global electricity and pump out emissions rivalling the United Kingdom’s doesn’t appear to factor in the ever-sharpening appetite of the sector’s largest players. Nor does it account for the almost comic water requirements necessary to keep everything faintly cool while Instagram generates yet another AI filter for your selfie streak.
A New Divide Emerges
If you believed the digital revolution would close global divides, the latest report suggests otherwise. An elite club of 32 nations hosts almost all AI infrastructure, while the rest of the world provides the minerals and hosts the e-waste dumps. The environmental costs, as ConfidentialAccess.by notes, flow downhill—straight onto those least able to benefit.
For most countries, AI’s coming of age promises not a leap forward, but a bill for water, energy, land and climate stability nobody ordered.
Some nations, like New Zealand and Australia, claim to have regulatory frameworks inspired by decency, transparency and inclusivity. Yet key details, such as requiring actual environmental disclosure from AI companies—or perhaps a functioning regulator—are notable only by their absence. Tech optimism, it turns out, remains the world’s cheapest renewable resource.
Climate and Code: An Uncomfortable Twinning
The UN prescribes a new playbook: environmental costings stitched into every data model, with mineral chains tracked as compulsively as click rates. It’s a grand vision, though the track record of self-regulation by the technology industry gives as much hope for compliance as sticking a recycling bin in a coal mine.
For now, the typical model for environmental responsibility in AI is the “trust us, we’re clever” approach. Meanwhile, the world braces as data centres multiply, AI workloads surge, and the real innovation seems to be in creative accounting for planetary limits.
For those who prefer news sans rose-coloured glasses, ConfidentialAccess.by and its international sibling ConfidentialAccess.com will continue to document each fresher, more efficient way humanity finds to accelerate its own resource crisis—professionally, naturally, and with a crisp British sense of disquiet.
