Artificial intelligence (AI) is everywhere these days. From chatbots like Siri and Alexa to self-
driving cars and personalized recommendations on Netflix, AI is changing the way we live and work. But as AI continues to grow and become more advanced, there's an important question we need to ask: What is the environmental cost of powering all this technology?
While we don’t yet have all the data to fully measure the environmental costs of AI, we do know that it takes up a huge amount of electricity, and it also uses a lot of water to keep things cool - making its environmental impact a serious concern.
As AI grows, these environmental costs will continue to get even bigger, if we don’t keep ourselves in check!
In this Need to Know News, we’ll take a look at how AI affects the environment, focusing on its energy use and its impact on water resources, and explore how we can reduce these costs and make AI more sustainable as it continues to evolve.
Let's dive in!
The Huge Energy Demands of AI
Training AI models—especially large systems like the ones behind popular AI chatbots—takes a lot of computational power. To train these models, we need massive servers that run 24/7, which consume large amounts of electricity. By 2026, electricity consumption from data centres and AI could reach up to 4% of global electricity demand, which would be double the amount recorded in 2022.
According to the United Nations Environment Programme (UNEP), training one large AI model generates around 300 000 kilograms of carbon dioxide (CO₂). On average, a single mature tree can absorb about 22 kg of CO₂ annually. This means we would need over 13 600 trees to offset a large AI model’s yearly CO₂ emissions.
Now, imagine the thousands of AI models existing around the world today. We would need a LOT of trees to offset AI’s carbon dioxide emissions.
The demand for energy to run these models is linked to the growing number of data centres around the world. Data centres are facilities where all the data is stored and processed, and they’re the backbone of the Internet and AI.
Over the past decade, the number of data centres worldwide has skyrocketed from 500,000 in 2012 to more than 8 million. By the end of 2030, the data centre industry will see approximately 2.5 metric billion tons of CO₂ produced.
As AI continues to grow, energy demand will also increase. Energy consumption from data centres is projected to double every four years, significantly driven by AI.
Excessive Water Consumption: The Hidden Cost of AI
Energy isn’t the only environmental resource AI is using. There’s also a significant water cost involved, and it’s not something we often think about when talking about technology. Data centres rely on large amounts of freshwater to cool their servers.
Without cooling, the machines overheat and cause massive power outages – and we all know how much of a pain it can be to deal with these outages.
While some data centres use air conditioning to keep things cool, many of the most efficient centres use liquid cooling systems, which require a lot of water.
Here are some issues that come with AI’s water consumption:
Evaporation loss: Cooling towers often use water that evaporates during the cooling process, leading to significant water loss over time.
Thermal Pollution: The heated water discharged from cooling systems can increase the temperature of nearby water bodies, potentially harming aquatic systems.
Chemical Contamination: Water used in cooling systems may contain treatment chemicals to prevent scaling or microbial growth. Discharge of treated water into natural systems can cause environmental contamination.
Drought Risks: In drought-prone areas, reliance on water cooling systems can exacerbate water shortages and raise public concerns.
In regions where water is already scarce, this can be a huge problem. Water is essential for drinking, agriculture, and ecosystems. The demand for it is only increasing due to climate change and population growth. As more and more data centres pop up to support AI and other tech, these centres compete for the same limited water supplies that local communities depend on.
For example, data centres in warmer climates need more water to cool their equipment. In parts of Southeast Asia, where water shortages are already an issue, this drives competition for fresh water to get even worse. As AI and other tech industries grow, they may clash with the need for water for agriculture and drinking—especially in areas where every drop counts.
What Can We Do to Make AI More Sustainable?
Fortunately, there are ways we can make AI more sustainable. Here are a few things that can help reduce the environmental impact of AI:
Switch to renewable energy. Switching to renewable energy, such as wind or solar energy, to power data centres will help reduce the environmental impact of AI.
Improve Energy Efficiency. Data centres can become more energy-efficient by using technologies such as carbon-efficient hardware and smaller AI models that require less energy to reduce their electricity consumption.
Water Recycling. Some data centres are starting to use closed-loop water recycling systems to reuse water for cooling instead of pulling in new water from local sources. This would help reduce the strain on water resources.
Conclusion: Balancing Innovation with Sustainability
AI is a powerful technology that has the potential to change the world in countless ways, from monitoring the environment to helping businesses make greener choices.
But as AI continues to grow, we need to be aware of its environmental impact. Its massive energy consumption and growing demand for water are challenges we can’t ignore.
By focusing on more sustainable practices—like using renewable energy, improving energy efficiency, and recycling water—we can reduce the environmental costs of AI and ensure that this technology doesn’t harm the planet in the name of technological progress. We’re still in the early stages of understanding how AI affects the environment, but the good news is that there’s a lot we can do to make AI more eco-friendly as it continues to evolve.
The key is to find a balance between AI’s benefits and its environmental costs. By making smarter choices and leveraging technology to address some of these issues, we can steer AI toward a role that supports the planet, rather than exacerbates its problems.
Want to stay in the loop on our latest updates? Looking to get involved? Sign up here to join our PowerSeekers community today and create a more compassionate world with us!
