Artificial Intelligence (AI) is rapidly transforming industries, with global adoption reaching 72% in 2024. However, this expansion coincides with escalating climate concerns. The environmental footprint of digitalization is a pressing concern as this digital transformation is taking place in parallel with the depletion of raw materials, water stress, pollution and waste generation. There is an ongoing climate crisis with 2024 being declared the warmest year on record with the occurrence of extreme climate events. The environmental footprint of AI spans its entire life cycle, hardware production, software development, training, inference, and disposal, contributing to greenhouse gas emissions, water consumption, raw material depletion, and e-waste.
Data centres, central to AI operations, are highly resource-intensive, consuming vast amounts of energy and water. Projections indicate that by 2026, global AI energy demand could surpass the annual electricity use of Belgium, while water demand may reach 6.6 billion cubic metres, posing serious challenges in already drought-stricken regions like Chile and Arizona. The hardware behind AI, GPUs and chips, relies on minerals such as cobalt, lithium, and copper, often extracted under environmentally destructive and exploitative conditions, particularly in the Democratic Republic of the Congo. This raises issues of inequity, as communities least responsible for climate change disproportionately bear its burdens.
A holistic, sustainable approach is essential to ensure AI innovation supports economic growth without deepening the climate crisis.
