SUSTAINABILITY’S AI PARADOX Can AI serve as a tool for advancing sustainability or will it become an obstacle? IMD’s Julia Binder and José Parra Moyano elaborate on how organisations must strike a delicate balance between innovation and responsibility to harness the technology’s power for good
T he role of AI in sustainability is double-edged. On the one hand, the technology has the potential to drive solutions that accelerate the shift to a sustainable future. On the other, its own environmental and social impacts must be carefully managed, enabling businesses to mitigate the challenges while maximising the benefits. AI technology, while transformative, also carries a substantial environmental and social footprint. The unparalleled energy and resources required to power AI systems pose a significant risk to sustainability goals – below are some points of note. ENERGY CONSUMPTION Enormous computational resources are required to train AI models. Deep learning models, in particular, need large amounts of electricity to process vast datasets. This energy consumption significantly contributes to the carbon footprint of AI systems. For example, Microsoft’s carbon footprint increased 30 per cent in 2023 compared with 2020 due to indirect emissions from the construction of data-hosting facilities. However, from a computational point of view, once AI models are trained, their operational phase can be far more energy-efficient. Generating content with a trained model, for example, requires
much less computational power than the training phase. Nonetheless, the overall environmental impact remains high if businesses do not address the energy consumption used during training. WATER CONSUMPTION Data centres require powerful cooling systems to prevent servers from overheating, often relying on large quantities of water. In water- scarce regions, this can exacerbate environmental strain, especially as demand for AI-powered services grows. While companies such as Google have made efforts to reduce water usage through more efficient cooling technologies, the broader industry still relies heavily on water‑intensive processes. RESOURCE DEMANDS AI’s demand for cutting-edge hardware is accelerating the depletion of finite resources. The specialised processors and memory units used in AI systems require rare earth elements, metals “AI technology carries a substantial environmental and social footprint”
and other resources. The mining and extraction of these materials contribute to environmental degradation, including pollution and deforestation. Furthermore, as AI hardware evolves rapidly, older devices quickly become obsolete, generating significant amounts of e-waste, which can harm the environment if not properly recycled. JOB DISPLACEMENT The advancement of AI also threatens job security, particularly in sectors that rely on routine tasks. According to the World Economic Forum, jobs in clerical and secretarial work are highly susceptible to automation, as AI can easily handle tasks like data entry, scheduling and customer service. While AI has the potential to create new job opportunities in emerging tech fields, the displacement of workers in more vulnerable sectors raises concerns about economic inequality and social stability. Upskilling and reskilling initiatives are essential to help workers adapt to the changing landscape and ensure that the transition to an AI-powered economy is equitable. However, AI offers remarkable potential to accelerate sustainability in spite of these challenges. By optimising resource use, enhancing data reporting and enabling innovative business models, AI can be a powerful ally in the fight for a sustainable future.
36 Business Impact • ISSUE 2 • 2025
Made with FlippingBook - Share PDF online