UN report warns AI could soon use 3% of world’s electricity and more water than we need to drink - Amanda Turnbull-McRae

A UN report highlights AI’s escalating environmental impact, projecting that by 2030, AI systems could use 3% of global electricity—equivalent to doubling current data center consumption—and require 9.3 trillion liters of water, an area nearly 10 times the size of Mexico City. The carbon footprint from this demand would need 6.7 billion trees grown over a decade to offset, while data centers already consumed as much electricity as Saudi Arabia in 2023. The report also exposes structural inequities, with just 32 nations hosting AI-specific cloud infrastructure, 90% of which is located in the US and China, leaving other countries to bear disproportionate environmental costs from mineral extraction and e-waste. The report emphasizes that AI’s operational footprint depends on both usage volume and efficiency, with different tasks—such as text, code, image, or video generation—requiring varying levels of computational effort. It argues for full value-chain governance, from mineral sourcing to recycling, and urges environmental disclosures to become standard in AI development. The goal is to balance AI’s capabilities with environmental stewardship, integrating projected AI demand into climate and energy planning. While governments globally promote AI adoption, few address its sustainability risks. Aotearoa New Zealand’s national AI strategy, informed by OECD principles, lacks mandatory environmental disclosures or energy-use tracking despite its focus on inclusive development. Similarly, Australia’s AI plan, which includes tools like Bowerbird for audio transcription and AI-assisted veteran claims processing, adopts a principles-based approach without enforcing sustainability measures. Both countries’ ‘light-touch’ regulation risks overlooking AI’s growing environmental cost, which cannot be mitigated solely by efficiency improvements. The report warns that without intervention, AI’s expansion could worsen resource depletion and carbon emissions, threatening economies and cultures dependent on natural environments. It calls for a shift in the AI innovation paradigm, prioritizing sustainability alongside technological advancement. The absence of regulatory frameworks to monitor AI’s energy and emissions impacts could undermine global efforts to address climate change, making responsible governance essential for a sustainable tech future.