They weren’t lying when they said technology is going to bring the world to our fingertips, but as much as we wish no one was lying, as they claimed that humans would bring about our own destruction. Through decades, we’ve caused the extinction of nearly 100,000 animals per year, deforested lands, and torn apart the ozone layer—yet we refuse to stop.

It's become common practice to make AI do the simplest of tasks, such as writing emails, but when that “send” button is hit, no one considers the consequences of their actions. It's kinda like a message, right? No, it's spurring climate change concerns.

According to the International Energy Agency Energy and AI report, the electricity demands from data centers will be around 950 TWh— to put it into perspective, that is slightly more than the energy consumption of Japan. The most significant driver of this increase is AI-optimised data centers. These centers have been around since 1945, but the speed at which AI consumes energy is the main concern. Postdoc in the Computer Science and Artificial Intelligence Laboratory (CSAIL), Noman Bashir states, “What is different about generative AI is the power density it requires. 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 main point being that AI is quickly depleting resources, and despite the majority of data centers being the heart of the problem, the rate at which users input commands and rely on AI expands the issue. 

Researchers have estimated that a single ChatGPT query by a user who needs an email summarized could amount to five times more electricity than a simple web search. Bashir points out that a major flaw is the lack of knowledge surrounding the environmental impact of their actions, which leads users to hold no incentive to attempt to curb their usage, further contributing to the crisis. Concerningly, ChatGPT users send 2.5 billion prompts a day, with many blissfully unaware of the damage caused. Many of these prompts are simply things that could be learned from a five-minute search or writing a couple of sentences, leading to a complete waste of energy on tasks that didn’t necessarily require it. 

Beyond the impacts it has on energy usage, AI data centers use up to five million gallons per day and 1.8 billion per year, which is equivalent to water usage from a town of 50,000 people. This is often used to cool the data centers, which grow burning hot from energy consumption. This reduces the availability of water for areas that may already struggle to find resources. Within places like Silicon Valley, data centers are pushing out farmers due to a lack of resources, and increasing taxes—all while these tech giants get tax breaks. 

Along with the negative impacts on water, these data centers are fueled using fossil fuels. Fossil fuels release tons of greenhouse gas emissions, which drive up temperatures and are a key factor in global warming. 56% of the energy needed to fuel data centers comes from fossil fuels; however, this also adds to the issue of water usage. Fossil fuel-generated power plants rely on large boilers full of water to be heated by natural gas or coal and then produce steam to turn a turbine. Ultimately, this adds extra stress, especially in areas that are susceptible to drought in the summer. When a metadata center opened up in Newton County, Georgia, a couple’s home was unable to run water through its taps, and the county is predicted to be in a water deficit by 2030. AI is taking away resources from regular citizens, leaving them stranded in the world of big tech vs small people.

All in all, while social media and the technology we have today are oftentimes not produced by the most ethical means, and may have a similar effect on the environment when compared to AI, AI is making a significantly higher impact in diminishing resources and propagating greenhouse gas emissions. It is extremely important as a user to cut back on the number of queries we put into generative AI and to avoid using it as much as possible. 

Additionally, it is important to appeal to legislators and politicians in order to make sure that data centers are not overtaking our communities, nor are they receiving tax breaks that ultimately citizens have to pay for. 

Kevin Park, a Santa Clara city councilman, said, “If you ask how many more data centers we need, the answer is always more”. A simple statement, but in reality a threat.