An MIT study estimates that by 2026 electrical consumption of AI data centers will have increased tenfold since 2022 – this increase makes AI data centers the largest consumer of electricity, consuming more than the entire country of France.
Though people use artificial intelligence daily few people are aware of the environmental impact and damage it causes to the natural world. Technology is not directly associated with being harmful to the environment as to most people in school environments it has a negative connotation with cheating. In order to uncover how AI harms the environment, the how and why must be explained.
“AI’s effect on the environmental world is hidden in 3 main categories: electricity, freshwater usage and rare earth metals”, Ms. Tordillos explained.
The source of energy used for AI operations is largely influenced by the location of the data centers, as different states rely on varying energy sources. For example, states like Washington utilize hydropower, a renewable energy source, while others may depend on non-renewable and environmentally harmful sources such as coal, oil, or nuclear energy.
“The amount of water AI utilizes is unexpected to many, but the reason this resource is so largely used is because the AI data centers need a lot of cooling power to keep the machines that process AI data from overheating: this is done through a heat exchange system, where water pipes help cool down the equipment”, said Tordillos when describing the use water has regarding AI data centers.
Despite the environmental impact is leaves behind artificial intelligence has made huge advancements in almost every field of work thus far, making the topic of AI a double-edged sword. Though many positive outcomes have come from AI, it does not take away all the damage it has consequently caused.
Tordillos points out that the biggest environmental problem with AI right now is its open access—where anyone can use it however they want, with no rules or limits. She believes setting boundaries to regulate how AI is used every day could help lessen its impact on the environment. Ultimately, Tordillos says, “We could make it more accessible once we have the right sustainable practices in place.”
These ‘sustainable pairings’ Tordillos mentions refers to the idea of ensuring that all the electricity used to fuel AI comes from solar energy or wind and not coal, nuclear or oil energy plants. Furthermore, with freshwater being such a scarce source in many places attempting to reuse or recycle water used in the cooling system instead of constantly using freshwater could also help reduce waste of such important environmental sources. Tordillos mentions the ideas of, rain collection systems, reverse osmosis and the use of gray water as alternatives to the current overconsumption of fresh drinking water used to cool down machinery in AI data centers.
“Being thoughtful about where these data centers are located would greatly help as well so that we are locating them where there is already access to clean energy and or access to a water source”, Tordillos
“The Industrial Revolution, marked by the rampant burning of fuels, drastically altered the composition of our atmosphere”, Tordillos relates the environmental impact of the Industrial Revolution to the to the destruction AI inflicts our environment with currently.
Therefore, it is important to recognize that although AI has transformed various industries, its environmental impact—through energy use and water consumption—is detrimental to the environment. Tordillos emphasizes that regulating AI’s use and adopting sustainable practices, like renewable energy and water recycling, could reduce its harm. As with past technological revolutions, it’s crucial to balance progress with environmental responsibility to ensure a sustainable future.
