How AI is Destroying our Environment

On November 30th, 2022, The servers for Chat GPT went online. Since then, the AI model has become the first to be used so widely and regularly. AI chatbots are used in schools, professions, personal life, and more. Since its inception, fears of what negative impacts high-level AI might bring have sprung up. Most recently, the environmental impacts of AI have been brought to the forefront of the conversation.

The Energy Demands of AI

Complex AI models require significant amounts of energy to both train and run them. With AI being used by millions of people every day, that level of energy consumption increases rapidly. Research by PhD candidate Alex DeVries revealed that Google's “AI Overview” chatbot uses 10 times the amount of energy as a usual Google search, and, by 2027, the annual energy consumption of Google’s AI chatbot could match that of the entire country of Ireland.

In order to match the newly found demands of energy levels from AI usage, energy suppliers will have to ramp up production. This means more fossil fuels will be burned, releasing more carbon dioxide into the atmosphere, contributing to the greenhouse effect, and increasing the global temperature.

Photo by Chris LeBoutillier on Unsplash

“What is different about generative AI is the power density it requires,” said Norman Bashir, an MIT Climate and Sustainability fellow in an MIT News article. “Fundamentally, it is just computing, but a generative AI training cluster might consume seven or eight times more energy than a typical computing workload.”

The article also helped quantify the drastic impact the introduction of AI has already made in the energy sector. At the end of 2022, the power requirements of data centers in North America stood at 2,688 megawatts. No more than a year later, that number had practically doubled to 5,341 megawatts by the end of 2023. The sudden and drastic increase in energy usage is largely due to the spreading of mainstream AI, which has become freely usable and available to people everywhere.

AI is Running Us Dry

Water is a vital resource for sustaining life. An abundant 70% of our planet's surface is covered with water. Despite that, only roughly 3% of the planet's water is freshwater that is safe for consumption by living things. That 3% gets even smaller when you realize a little over two-thirds of that 3% is locked away in polar ice caps that we can't access. The scarcity of water combined with its vital importance for all living things on the planet makes it all the more alarming when you see just how much of this scarce resource is being used to continue running AI models.

AI models are powered and run by huge data centers that withhold the machines that allow AI to function. Because of the massive amount of energy that AI requires, these data centers quickly and very easily overheat. When the data centers begin to overheat, water is applied, cooling down the servers. To give you a scope of how much water data centers use on average, the data centers used by Google to power their cloud system, Gmail, Google Drive, and YouTube consume roughly 550,000 gallons of water a day. That equates to 200 million gallons of water a year.

The extensive and wasteful usage of water extends into two main different consequences. 

First, it can lead to scarcity. As explained earlier, only a very small amount of the earth’s water is available for human consumption, and many places around the world already struggle to obtain sufficient amounts. As AI grows and more data centers are built that require huge amounts of water, the resources will become even more scarce, leading to malnourishment and dehydration among increasingly more communities worldwide.

Secondly, the water used at data centers to cool them off evaporates when coming in contact with the servers. Increased evaporation around the world is a contributing factor to the frequency and severity of storms. The water cycle explains how once water evaporates into the atmosphere it eventually comes back down through rainstorms. If water is entering the atmosphere more frequently and at higher volumes through these data centers, it means that storms will happen more often and at greater magnitudes. This will eventually lead to more deadly hurricanes and thunderstorms that can decimate coastal regions as they have in the southeastern United States in recent years.

How Companies are Using AI for the Worse

The impacts of AI extend beyond the emissions they produce too. Cooperations that produce significant amounts of pollution, carbon dioxide emissions, and environmental damage are using AI to grow their businesses. To see this, we can first take a look at ExxonMobil, a huge oil and gas company. In 2019, ExxonMobil announced a partnership with Microsoft that would lead them to work together to train and develop AI models that revolutionize their gas company’s efficiency. Through their cooperation, ExxonMobil believed they would be able to increase their production by roughly 50,000 barrels of crude oil per day. The increase in production results in greenhouse gas emissions equivalent to what 4,300 gas-powered cars produce in a year, according to the EPA's emissions calculator.

Photo by Zbynek Burival on Unsplash

The oil industry isn’t the only sector benefitting from the use of AI. Fast Fashion companies are utilizing AI algorithms to improve their advertising strategies, leading to dramatic increases in sales across the globe. As those sales grow, people will play into patterns of overconsumption, buying things they don't really need. The products they buy all require finite resources to make, and the production of them can cause admissions from factories and the materials used in fast fashion clothes often shed microplastics into the environment.

“Without strong ethical, social, and environmental standards in place, AI could just as easily be driving faster production and overconsumption,” said Lewis Perkins, the president of the Apparel Impact Institute, a global nonprofit that measures the fashion industry’s climate impact, in Time Magazine.

Among fast fashion companies around the world, Shein has established itself as one of the most well-known and highest-grossing brands. With their success, has come huge environmental impacts. The company has explained how they have developed AI models and algorithms that specialize in optimizing sales. Because of their recent utilization of AI, the company’s carbon dioxide emissions have reached a total of 16.7 million metric tons in 2023 according to their most recent sustainability report. That number is so high that it is double the company's emissions from 2022, and earned them the title of the world's biggest polluter in the fashion industry.

“AI enables fast fashion to become the ultra-fast fashion industry, Shein and Temu being the fore-leaders of this,” said Sage Lenier, the executive director of Sustainable and Just Future, a climate nonprofit. “They quite literally could not exist without AI.”

Keeping up with AI

The underlying issue with AI and its impacts is just how fast the technology is developing. Within the last 5 years, AI models have evolved at an unprecedented pace. So fast, that regulations have been unable to keep up in limiting the rampant growth of this sector. There are logical solutions to all the problems that generative AI is spawning, however, it has been seemingly impossible to find those solutions because of how fast the technology is developing. As it stands, we are letting this form of technology get out of our hands, so much so that its negative impacts could one day come back to hurt us in a bigger way than we could have ever imagined.