The environmental costs of artificial intelligence threaten water, land, and climate

Data centers running artificial intelligence systems could consume up to 945 terawatt-hours of electricity annually by 2030, which is three times the total annual electricity consumption of Pakistan, Bangladesh and Nigeria combined, countries with a population of more than 650 million.

However, this is just the tip of the iceberg. In addition to the carbon footprint, every unit of electricity consumed by data centers involves a "water footprint" for cooling and power generation, and a "land footprint" related to power generation and supply chains.

Reconsidering the sustainability scale
According to a new study by the United Nations University, AI-related water consumption could equal the basic annual household needs of 1.3 billion people by the end of this decade, while its land footprint could exceed 14,500 square kilometers, which is about five times the area of ​​Cairo Governorate.

The report highlights a fundamental gap in how the environmental impact of artificial intelligence is measured, as the focus is often on greenhouse gas emissions, particularly those associated with training large models, while this approach overlooks other environmental costs.

Solutions that are considered "green" from one perspective may exacerbate pressures in other areas, particularly in regions already suffering from resource scarcity. For example, switching to certain renewable energy sources may reduce carbon emissions, but it could also lead to a significant increase in water consumption and land use.

Daily use of artificial intelligence
Public debate has largely focused on the energy required to train advanced AI models, but the study showed that daily usage accounts for between 80 and 90 percent of total energy demand, and the figures are staggering.

It is estimated that one widely used artificial intelligence service processes about 2.5 billion directives per day, consuming hundreds of gigawatt-hours of electricity annually.

Energy consumption also varies significantly depending on the task. Generating a single image using artificial intelligence may require more than a thousand times more energy than classifying a simple text, while generating videos requires much greater resources.

Improvements in efficiency alone are unlikely to offset this increased demand. The report pointed to a known feedback loop, where lower costs and improved performance lead to higher utilization rates, ultimately resulting in an overall increase in resource consumption.

Local burdens and global benefits
The environmental impacts of AI infrastructure are not evenly distributed. While the benefits of this technology are global, its costs are often concentrated in specific regions.

In some countries, data centers already account for a significant share of national electricity consumption, putting strain on power grids. In others, expanding facilities are significantly depleting water resources, even during periods of drought.

At the same time, the report warned of a growing challenge in the form of electronic waste, with AI infrastructure expected to generate up to 2.5 million tons of e-waste annually by 2030. Much of this burden is likely to fall on low-income countries that lack sufficient capacity to safely dispose of this waste.

The production of essential minerals needed for artificial intelligence devices also raises concerns about environmental degradation and social injustice in the areas where extraction takes place.

A widening digital and environmental gap
The expansion of AI infrastructure is also creating new inequalities in access and influence. According to the report, more than 90 percent of the world’s AI computing power is concentrated in just two countries: the United States and China. In contrast, more than 150 countries lack a substantial domestic AI infrastructure.

This imbalance not only limits economic opportunities, but also raises questions about environmental justice, as some countries bear the environmental costs without reaping the benefits of AI-driven growth.

Towards responsible artificial intelligence
Despite these stark findings, researchers at the United Nations University stressed that the report is not intended to oppose artificial intelligence per se, but rather calls for urgent action to ensure that this technology develops within the planet's environmental limits.

The study proposed a framework for a "responsible AI ecosystem," based on principles including transparency, efficiency from the design stage, fairness, accountability throughout the lifecycle, global collaboration, and sustainable use.

Governments were urged to integrate AI infrastructure into energy, water, and land-use plans, and companies were encouraged to design systems that reduce resource consumption. Users also have a role to play, by choosing applications with the lowest environmental impact whenever possible.

In conclusion, the report concluded that the future of artificial intelligence will depend not only on technological innovation, but also on the governance choices made today.