AI’s deepening impact on cities, climate and global labor justice

A groundbreaking study has spotlighted the rapidly expanding ecological and labor footprint of artificial intelligence (AI), revealing how this technology is reshaping urban landscapes, resource flows, and societal inequities. The study, titled “The Nature of AI: Metabolism, Energy, Water, Labour and Justice in the Urban Political Ecology of Artificial Intelligence”, was published in the journal Urban Political Ecology.

Authored by scholars from the University of Manchester and King’s College London, the paper critically examines AI’s role through the lens of urban political ecology, focusing on the hidden material and social infrastructures that power intelligent systems .

Through an in-depth conceptual framework and empirical cases from North America and Europe, the study reveals that AI’s development is not a disembodied digital process but a deeply physical and geographically rooted one. The research demonstrates how data centers, energy sources, and labor networks are being reorganized to support the computational metabolism of AI, with cities playing a pivotal role in this transformation.

How is AI altering urban energy and resource flows?

The study introduces the concept of AI metabolism to explain how intelligent technologies consume urban energy, water, and labor in specific and escalating ways. Data centers, the physical backbone of AI, are positioned at the heart of this analysis. These facilities require continuous power, water for cooling, and complex digital infrastructures to support the training and deployment of AI models. Cities, particularly in the Global North, are becoming key territories in this arrangement as they host hyperscale data centers and smart infrastructure.

Case studies from Dublin and Virginia illustrate how urban planning and policy are being reoriented to accommodate AI’s infrastructural needs. In Dublin, for instance, data centers now consume nearly 20% of Ireland’s electricity, a figure expected to rise as AI systems scale. Similar concerns are raised in the United States, where AI-linked server farms are drawing increasing scrutiny for their environmental impact.

The authors argue that this material expansion of AI is driving cities into deeper metabolic dependencies on fossil energy, rare-earth minerals, and water-intensive processes. The result is a technological metabolism that not only consumes significant resources but also reshapes energy policy, utility pricing, and environmental planning at the municipal level.

Who builds AI and at what cost to labor and justice?

Beyond infrastructure and energy, the paper addresses the human labor embedded in AI’s operation—particularly the invisible workforce behind data annotation, content moderation, and system training. These workers, often based in the Global South, form what the authors describe as the “hidden labor” of AI: underpaid, precariously employed, and essential to the system’s functionality.

The study highlights how this labor is often excluded from policy discourse around AI ethics, which tends to focus on algorithmic fairness or bias without considering the working conditions of those who make AI possible. Workers in Kenya, the Philippines, and Venezuela were cited as integral to major platforms’ success, yet receive little protection or recognition. In contrast, tech labor in the Global North is increasingly centralized in “AI urban clusters” like San Francisco, London, and Berlin - areas that benefit from concentrated investment, education, and political capital.

This divergence produces a two-tiered geography of AI labor: one elite and high-income, the other exploitative and invisible. The researchers frame this as a new form of global inequality that mirrors historical patterns of colonial extraction, only now applied to cognitive and emotional labor outsourced across digital platforms.

What should policymakers and planners do about AI’s urban impact?

The study concludes with a series of normative and policy-oriented reflections. First, it calls for a fundamental rethinking of how cities govern AI’s infrastructure. This includes embedding environmental regulations into tech zoning laws, introducing labor protections for platform workers, and increasing transparency around AI’s energy use and carbon footprint.

Second, the researchers propose a shift in how AI development is conceptualized in urban planning. Rather than treating AI as a weightless, invisible tool, policymakers should recognize it as a high-impact urban actor, one that requires roads, power, water, housing, and human oversight. By integrating AI infrastructure into climate policy, social equity frameworks, and digital rights agendas, cities can better manage its complex impacts.

Lastly, the paper urges scholars and technologists to adopt a political ecology perspective that foregrounds justice and materiality. This means expanding AI ethics beyond algorithms to include energy justice, labor equity, and environmental sustainability. Cities must play a central role in this realignment, acting as laboratories of governance where AI’s costs and benefits are most acutely experienced.