Artificial intelligence is simultaneously being promoted as a solution to environmental challenges and identified as a significant driver of ecological degradation. According to Dr. Yannis Dafermos, an academic at the School of Oriental and African Studies (SOAS), the rapid expansion of AI infrastructure is exacerbating the environmental crisis rather than alleviating it, despite widespread techno-optimist claims that AI will help solve climate problems. Dr. Yannis Dafermos Is AI Infrastructure Actually Making Environmental Problems Worse? The physical infrastructure required to power AI systems demands enormous amounts of electricity and water. In many regions, this electricity is still generated primarily from fossil fuels, while data center cooling consumes vast quantities of water. This creates a fundamental tension: the technology being marketed as an environmental solution is itself becoming an environmental burden. The concern extends beyond energy consumption. AI infrastructure damages local ecosystems and affects the livelihoods of communities that depend on them. These localized harms cannot be compensated for by efficiency gains achieved elsewhere in the world, according to Dr. Dafermos's analysis. Dr. Dafermos "The techno-optimist view that AI can solve the environmental crisis is highly problematic: it acts as an excuse to delay the systemic transformations that are essential," explained Dr. Yannis Dafermos, an academic at SOAS. Dr. Yannis Dafermos, Academic at School of Oriental and African Studies While AI technologies do offer genuine environmental benefits, the question remains whether these gains will be sufficient to offset the growing footprint of AI infrastructure itself. AI systems can improve energy efficiency in buildings, optimize heating systems, manage electricity grids more efficiently, reduce waste in supply chains, and enable precision agriculture that uses less water and fertilizer. However, the uncertainty about whether these benefits will outweigh the costs is substantial. How to Evaluate AI's True Environmental Impact - Assess Infrastructure Costs: Calculate the total energy and water consumption required to build and operate AI data centers, including the carbon intensity of the electricity grid powering them. - Measure Offset Gains: Quantify the actual environmental improvements achieved through AI applications in energy management, agriculture, and supply chain optimization. - Account for Local Damage: Evaluate the ecological and community impacts of data center construction and operation in specific regions, not just global efficiency metrics. - Challenge Techno-Optimism: Question whether technological solutions alone can address systemic environmental challenges without fundamental economic and social transformation. Why the AI Bubble Matters Beyond Tech Investors The financial dynamics surrounding AI investment carry risks that extend far beyond Silicon Valley. While mega-corporations like Alphabet, Amazon, Meta, Microsoft, and NVIDIA have substantial profits to fund AI development, the broader AI ecosystem relies increasingly on debt and shadow banking mechanisms. This includes data center developers, AI startups, and utility companies expanding electricity generation to meet growing energy demands. These financially fragile companies in the AI supply chain are more vulnerable to defaults if anticipated profits fail to materialize. The parallels to the dot-com bubble are striking: euphoric expectations about transformative technology, stock price overvaluations without corresponding profits, and investment levels exceeding actual demand. However, the current situation differs in that today's technology leaders are already highly profitable mega-corporations with resources to sustain investment. The critical question is whether AI technologies will be adopted widely enough in the economy to generate productivity gains matching investor expectations. If adoption falls short, stock prices could decline and financial losses could cascade through the AI ecosystem, similar to what happened during the dot-com crash. The Labor Question: Should Society Organize Around AI or AI Around Society? Beyond environmental and financial concerns lies a deeper question about how AI should shape society. Dr. Dafermos argues that visions centered entirely around AI technology, while ignoring political, social, ethical, and environmental issues, are fundamentally flawed. The risk is concentrating unprecedented economic and political power in the hands of mega-corporations claiming they can deliver an AI-driven future. "Our vision should not be to organise societies around technology, but to shape technology around democratically defined social and environmental priorities," stated Dr. Yannis Dafermos. Dr. Yannis Dafermos, Academic at School of Oriental and African Studies Large-scale robotics and AI systems required for a fully automated society would depend on vast amounts of energy, raw materials, and water, raising serious sustainability concerns. Additionally, these systems are likely to reproduce or amplify existing biases and forms of exploitation embedded in data and labor supply chains. The invisible workforce behind AI development illustrates this problem. A significant proportion of workers performing repetitive tasks essential for training machine learning models, such as image tagging and speech recording labeling, are based in the Global South. These workers operate under precarious conditions and receive extremely low wages. Rather than guaranteeing human flourishing, AI technologies risk deepening social injustice, economic precarity, and ecological degradation. The fundamental challenge is that techno-optimism about AI solving environmental and social problems may actually delay the systemic transformations necessary to build sustainable and equitable societies. Until AI development is explicitly shaped by democratically defined priorities focused on social justice, well-being, and environmental protection, the technology will likely continue to concentrate power and exacerbate existing inequalities.
