The Fourth Industrial Revolution is an era of technological transformation characterised by the convergence of digital, physical, and ecological systems. Artificial Intelligence sits at the heart of it, quietly reshaping every aspect of our lives. AI can process information and, at least superficially, mimic human thinking. It curates our social media feeds; manages logistics behind the products we buy and is increasingly running quietly in the background of almost every industry.  

 

But intelligence, it turns out, is not weightless. 

Behind every AI interaction lies a vast and energy-hungry infrastructure. It consumes enormous amounts of electricity, water, and physical hardware. These systems are powered by data centres running on fossil-fueled grids, cooled by freshwater and built from rare-earth minerals mined under extractive conditions.  

 

A single request made through an AI chatbot consumes up to 10 times more electricity than a standard internet search.  

The cloud, it seems, has a carbon shadow.  

A carbon footprint refers to the greenhouse gases released into the atmosphere as a result of human activities. Driving cars, flying planes, heating homes and now, using AI system all adds to this burden. 

The convenience of “thinking” machines comes at a planetary cost, one that is rarely disclosed to the user.  

 

When Intelligence Runs Thirsty 

 

Data centers use water to cool electrical components. Globally, AI infrastructure may soon consume six times more water than Denmark, a country of 6 million. This is a problem when a quarter of humanity already lacks access to clean water and sanitation. Most large-scale AI deployments are housed in data centers which need rare earth elements, which are often mined in environmentally destructive ways, not to mention the human rights violation there.  

These data centers also produce electronic waste, which often contains hazardous substances like mercury and lead.  

 

Green AI or Greenwashing? 

 

Governments and corporations are racing to develop AI strategies, but rarely do they take environment and sustainability into account. Despite climatic impacts, AI is frequently marketed as ‘Green’, ‘Carbon-neutral’, or ‘Sustainable.’ Many corporations emphasize efficiency gains while remaining silent on the full environmental costs of their infrastructure.  

In one striking example, Li et al. (2025) found that training GPT-3 at a Microsoft’s U.S data center consumed more than 5.4 million liters of water. Scaled globally, the impacts are staggering. Even everyday AI use adds up. A single GPT-3 query can consume a 500ml water bottle every 10 to 50 prompts, depending on where it’s hosted. 

 

The Global Price Tag 

 

Communities in the Global South and marginalized areas of the Global North bear the brunt of these carbon emissions, water withdrawals, e-waste and hidden labor that make AI possible (without sharing its benefits).  

As Kate Crawford writes in her book, The Atlas of AI, “What looks like machine intelligence is often just water, power, and labor in disguise.” 

Crawford states in one of her interviews, “I wanted to really open up this understanding of AI as neither artificial nor intelligent. It comes from the most material parts of the Earth’s crust and from human bodies laboring. It imposes a social order, naturalizes hierarchies, and magnifies inequalities.” 

Seen through this lens, AI can no longer be considered an objective or neutral technology.  

As Florian Zandt observes in his analysis of the 2023 Conflict Minerals Report, Amazon’s operations are implicated in mineral sourcing from African countries, such as the DRC and South Sudan, where armed groups leverage resource extraction to sustain violent conflict.  

These patterns echo long-standing legacies of environmental injustice, resource extraction, and digital colonialism.  

 

Digital Colonialism and Invisible Labor 

 

In her 2025 paper, Digital Colonialism, Samavia Zia argues that AI development replicated global power imbalance rooted in colonial history. Zia also exposes the hidden labor force behind AI. In countries like the Philippines, India and Kenya, underpaid workers perform critical tasks like flagging toxic content and moderating platform abuse.  

This so-called automation still depends on human labor but its labor that’s hidden and undervalued. AI surveillance tools are disproportionately deployed in postcolonial and marginalized communities, deepening existing structures of control. 

 

Individual Responsibility And Collective Reimagining 

 

Ironically, AI can also help solve the very climate problem it contributes to, but the question here is in the midst of a climate emergency, do we need ever-larger AI systems? Or do we need restraint, care, and accountability? 

Like Crawford states, “We have spent far too much time focusing on narrow tech fixes. Now we have to contend with the environmental footprint, with the very real forms of labor exploitation happening in these systems.” 

Many users are unaware that their playful queries contribute to energy consumption. By fostering greater awareness, we can encourage more thoughtful engagement. Users need to be nudged to consider if their query is worth the energy cost.  

Using AI with awareness is integral. Just as we learn to conserve electricity and water, we can learn to be more mindful of digital consumption.  

Treat AI as a tool and not a default. Consider the labor behind the screen. 

 

It is important for organizations to understand that value isn't limited to financial returns. Long-term value means protecting natural resources, fostering human well-being, and ensuring stable growth.  

We can challenge the idea that AI is abstract. We can ask the important questions, who builds it? Who benefits? Who bears the cost? What do we overlook? How might we build systems grounded in sustainability and care? 

No one wants to roll back the clock on breakthroughs, but if we continue ignoring the environmental costs, we risk creating a digital future that undermines the planet’s health, our ultimate source of life, and well-being. 

We owe it to ourselves and future generations, and the countless species with whom we share the world, to ensure that our digital dreams are not powered by depleted water and invisible labor.  

 

The future doesn't need smarter machines as much as it needs wiser choices.  

 

Image Credits:  Pexels