Artificial Intelligence, or AI, often seen as the future of technology, is changing everything from healthcare to finance. But as we get excited about its potential, we need to ask: Is AI sustainable?

Image generated with MetaAI

While it brings amazing efficiency and innovation, the environmental and social costs of AI can’t be ignored.

AI isn’t just a fancy term—it’s a technology that’s become a part of our daily lives. From smart assistants to personalized recommendations, AI’s impact is huge. However, beneath the cool algorithms and clever solutions lies a less attractive reality.

The energy use of AI systems, especially those for deep learning and big data processing, is enormous. A study by the University of Massachusetts Amherst in 2019 found that training a single AI model can produce as much carbon as five cars over their entire lifetimes.

Just the beginning

These systems require vast amounts of electricity, mainly because they run on powerful GPUs and TPUs (Tensor Processing Units). Data centres, the backbone of AI, are energy-intensive creatures.

According to a report by the International Energy Agency, data centres worldwide consume about 1% of global electricity demand, a figure that’s only expected to grow.

Moreover, the production of AI hardware is resource-heavy. Mining for rare earth metals, essential for creating these high-performance chips, has significant environmental impacts. The extraction process is not only energy-consuming but also leads to habitat destruction and pollution.

The e-waste generated when these components reach the end of their life cycle adds another layer to the environmental burden.
AI’s growing thirst for power, in turn, exacerbates water scarcity concerns worldwide.

Computational-intensive AI models, like OpenAI’s ChatGPT, housed in vast data centres, are significant culprits. Microsoft and Google have reported staggering water consumption increases, with Google using 5.6 billion gallons in 2022.

ChatGPT alone consumes about 500 millilitres of water for every 5 to 50 user prompts, underscoring the technology’s significant water footprint.

The consequences are tangible in areas hosting data centres like Iowa, US. Locals have seen escalating water rates and depleting water sources. The two tech giants, recognising their environmental responsibility, have pledged to mitigate these impacts.

Microsoft aims for 100% carbon-free energy and water-positive status by 2030. Similarly, Google is striving for net-zero emissions by 2030, emphasising the importance of sustainable AI development.

Then there’s the social impact

Beyond the environmental impact, AI has profound socio-economic consequences. One of the most pressing issues is job displacement. As AI and automation advance, many jobs, particularly those involving routine or repetitive tasks, are at risk.

A study by the World Economic Forum estimated that by 2025, 85 million jobs could be displaced by machines. While new jobs will be created, they often require skills that displaced workers may not possess. Fortunately, it’s the intrinsically ‘human’ jobs that will survive, and we will likely see an increase in focus on human-centric skills.

One of the most ironic aspects of AI’s sustainability issue is the energy paradox. AI can help optimise energy use in various sectors, potentially leading to more sustainable practices. For instance, AI can improve the efficiency of renewable energy sources, predict maintenance needs for machinery, and reduce waste in supply chains.

So the very technology that can help us be more sustainable is itself a major energy consumer.

AI is undoubtedly a powerful tool that can drive significant advancements across various fields. However, its sustainability is a critical issue that needs to be addressed. As we continue to develop and integrate AI into our lives, we must be mindful of its environmental and socio-economic impacts.