The United Nations University Institute for Water, Environment and Health has urged AI companies to disclose their environmental impact, focusing on energy consumption, carbon output, water usage, and land requirements.
Projected global data center electricity use will escalate from 448 TWh by 2025 to 945 TWh by 2030, triple the combined energy consumption of Pakistan, Bangladesh, and Nigeria. Water usage related to AI activities may reach 9.32 trillion liters, enough for 1.3 billion people in Sub-Saharan Africa. By the end of the decade, data center land use could exceed 14,500 square kilometers, larger than Montenegro.
The report emphasizes that carbon reduction does not guarantee lower water or land usage, highlighting the influence of local energy sources. A hydroelectric-powered data center may present favorable carbon metrics but could deplete water resources.
It establishes six principles for responsible AI development: transparency, efficiency, equity, lifecycle responsibility, global cooperation, and sustainable use, framing environmental justice as a central issue. AI’s environmental impact often burdens communities that least benefit from the technology.
The European Union is advancing mandatory sustainability reporting for tech companies, and this report could catalyze similar initiatives globally.