In recent years, AI has gained huge popularity in the fields of science, social studies, and technology for solving complex problems, from diagnosing diseases to forecasting floods. Its potential in addressing the climate crisis is also praiseworthy — for instance, in aiding environmental research, which has contributed to mapping deforestation and soil erosion, as well as forecasting droughts and wildfires. Activists are now using Global Fishing Watch, which employs machine-learning software to identify overfishing and illegal fishing in specific areas, thereby protecting our ecosystems. However, behind all this excitement, there is an uncomfortable truth — AI might be doing more harm than good for the environment.
AI’s Profit Problem
The rise of AI has opened many opportunities for businesses to maximise their profits, but this comes at a price. Businesses increasingly rely on AI, leading to increased resource extraction, industrial production, and global consumption. For instance, deep learning technologies are used by major corporations to analyse massive geological datasets, aiming to identify new oil and gas reserves. Additionally, smart submersibles, led by AI, can now detect untapped oil and gas reserves in remote parts of the ocean. Moreover, to generate profit, states and firms often exaggerate the benefits of new AI technology for personal gain, while downplaying or concealing its environmental costs. The tech world works by hyping people up — for example, when start-ups claim ‘technological breakthroughs’ to attract funding from investors. Large firms, such as Amazon and Google, dominate this space, controlling the technology and shaping public perception. The result? Increased production, overconsumption, and escalating pressure on the planet’s already limited resources.
AI’s Own Environmental Footprint
AI not only facilitates environmentally harmful practices but is also resource-intensive. Training advanced AI models, such as ChatGPT and GenAI, requires enormous amounts of electricity. According to the United Nations Environment Programme, a single search made through ChatGPT consumes ten times more energy than a simple Google search. Intuitively, complex models tend to be even more energy-intensive and have a negative impact on the overall environmental footprint.
Moreover, once an AI model is trained, the energy demand does not stop; rather, it continues to consume electricity every time it is used. The rise of AI has also fuelled the expansion of data centres, which require vast amounts of water to cool equipment and prevent overheating. They also produce electronic waste and generate significant greenhouse gas emissions. As companies constantly release newer and more powerful models, the carbon emissions continue to rise at a pace that far outweighs any emissions that AI tools could help reduce.
Smart Tech Needs Smart Governance
If left to corporations alone, AI risks being misused for short-term profits, leading to long-term environmental degradation. Without proper regulation, AI could become a tool of overproduction and unsustainable consumption. Hence, I believe that the best way to manage the risks of AI is to combine technology with politics to provide innovative solutions to environmental crises and to ensure that AI development is aligned with environmental and social goals.
The Government should more closely control AI development to prioritise its use for public interest, particularly among researchers, scientists, and institutions working on climate and sustainability. Additionally, state-funded AI initiatives can help transform industries with low-carbon technologies while also preventing businesses from overusing these technologies for purely profit-driven purposes. In today’s world, we are not passive receivers of technology; rather, we have the power to shape it to our advantage responsibly. Hence, we must redefine its purpose as a tool to ensure sustainable development.
(The author is a graduate in International Relations from the University of British Columbia, Vancouver)