United Nations has called on leading artificial intelligence companies to fully disclose the environmental footprint of their data centres and commit to powering their rapidly expanding infrastructure with renewable energy, warning that the AI revolution could become a major driver of global resource strain if left unchecked.
Launching a new AI Environmental Transparency Initiative during London Climate Action Week, UN Secretary-General António Guterres noted that the explosive growth of data centres powering artificial intelligence systems is already placing increasing pressure on global energy systems, water supplies and land use and that the world is not prepared for the scale of expansion projected in the coming years.
He urged companies developing and deploying advanced AI systems to move beyond voluntary climate pledges and adopt mandatory disclosure of their carbon emissions, water consumption and land-use impacts.
“By 2030, they could use more power than all but five countries and enough water to meet the basic needs of all 1.3 billion residents of sub-Saharan Africa for an entire year.”
António Guterres
António Guterres further added that, “if AI is to help build a better future, it must be honest about what it costs us now.”
The warning coincides with increased scrutiny of the hidden costs of AI infrastructure by environmental academics and civil society organisations, especially with regard to the enormous data centers needed to train and run large-scale models. Although AI is frequently discussed in terms of productivity, creativity, and economic revolution, its environmental impact has remained far less apparent.
AI Expansion Raises Questions Over Climate Equity and Global Governance
The rapid expansion of artificial intelligence infrastructure has created a complex environmental challenge that extends far beyond carbon emissions, exposing gaps in how the global community measures and regulates the technology’s impact.
According to UN findings current climate assessments of AI systems are incomplete. While emissions linked to training large AI models often receive attention, researchers argue that this focus overlooks the far larger environmental burden created by day-to-day usage.
The scope of the problem is considerably altered by this change in comprehension. An estimated 2.5 billion prompts are processed daily by one popular AI platform, using hundreds of gigawatt-hours of electricity every year. Depending on the application, energy requirements might also differ significantly. For example, creating AI images can take over a thousand times as much energy as basic text classification, and creating videos puts even more strain on systems.
Although significant, increases in computing efficiency could not be sufficient to counteract growing demand. According to the UN report, there is a rebound effect whereby more affordable and effective AI systems are used more frequently, ultimately increasing rather than decreasing overall energy consumption.
Beyond energy, the global impact of AI on the environment is not uniformly dispersed. The advantages of AI technology are worldwide, but the expenses are frequently concentrated in particular areas where data centers are situated.
Local grids are under strain in several nations where these facilities already make up a sizable portion of the country’s electricity use. In others, they largely rely on water supplies in regions that are already facing scarcity or drought. By forecasting that AI-related infrastructure might produce up to 2.5 million tonnes of e-waste annually by 2030, much of which is likely to be processed in lower-income nations with insufficient recycling capacity, the paper also warns of a developing electronic waste catastrophe.
These worries are exacerbated by the extraction of essential minerals needed for AI gear, which raises concerns about labour conditions and environmental deterioration in mining regions that supply the world’s technology industry.
In addition, there is a glaring global disparity in the distribution of AI infrastructure. The United States and China account for over 90% of the world’s AI-specialized processing capacity, whereas more than 150 nations lack substantial domestic AI infrastructure.
UN contends that a fundamental change in governance is necessary to address these issues. Guterres demanded that AI companies pledge to power all data centers with renewable energy by 2030 and to be completely transparent about the environmental effects of AI systems throughout their whole lifecycle.
In order to expedite what he referred to as a “just transition” away from fossil fuels, the Secretary-General also announced intentions to bring together world leaders prior to COP31 in Turkey, placing AI governance as part of a larger climate policy.
UN further emphasised that the report does not dismiss artificial intelligence per se, despite the cautions. Rather, it advocates for a more accountable and open development model that synchronises technological advancement with global limits.
READ ALSO: Rawlings Memorial Must Inspire Higher Governance Standards – Tsatsu Tsikata
