Welcome back to Lending a Helping Hand to Our Home! This month’s topic is one that I believe needs a lot more attention: the environmental impacts of generative AI. In recent years, generative AI has grown exponentially more popular; we use it to answer questions, entertain ourselves, and explore new topics that we are interested in. Most of us do not even think before we click on ChatGPT, no matter how we are using it. But, I am here to tell you that the next time you open your phone to use generative AI, you may want to be conscious of the impact your actions are having on the environment.
It may sound a little far-reached; after all, how is asking AI a question about a math problem really contributing to environmental problems? In reality, generative AI is bad for the environment in almost every way. Generative AI data centers use exorbitant amounts of electricity, produce countless tons of carbon dioxide, and require water for every stage of the process.
It makes sense that AI would need electricity, but according to MIT News, computational power required to just train AI uses a “staggering amount” of it. Now that popular Generative AI tools like ChatGPT and CharacterAI have been made available to the public for widespread use, the energy demands have only continued to increase. Per MIT, by 2026, the electricity consumption that data centers use is expected to grow so much that it will be fifth on the global list of electricity consumption, “between Japan and Russia.”
This new demand for electricity to power AI training centers and servers is creating other problems. According to Norman Bashir, a Computing and Climate Impact Fellow at MIT, “The demand for new data centers cannot be met in a sustainable way.” This means that in order to power our new AI habits, we are going to have to resort to using fossil fuels, something that is in direct conflict with the current shift the world needs to make away from fossil fuels. Another article released by MIT puts it best, saying the growth of generative AI, “runs counter to the massive efficiency gains that are needed to achieve net-zero greenhouse gas emissions by 2050.”
Water usage is another major concern regarding generative AI. According to the United Nations, around two billion people worldwide do not have access to safe drinking water. This means that we are arguably in the middle of a worldwide water crisis, and should be taking actions to make water usage more sustainable. But, as explained by Forbes, “AI’s projected water usage could hit 6.6 billion meters cubed by 2027.” That’s around 900 billion gallons, or close to the water usage of 11 million households a year. If we want to be sustainable about our water usage, and take action towards making sure all humans have access to the basic human right of clean water, AI cannot continue to grow as it is.
All of these problems seem like huge worldwide problems; after all, how much is your one conversation really contributing to the problems caused by AI? Well, per an article by UC Irvine, “a search driven by generative AI uses four to five times the energy of a conventional web search.” So, everytime you turn to AI instead of just Google or Safari, you are multiplying your energy usage by four or five. For every. Single. Question. According to the Yale School of Environment, someone who asks Generative AI 10-50 questions, “drives the consumption of a half-liter of fresh water.” So, even using generative AI only a few times has direct environmental impacts. And, students at PA use AI a lot more than that.
In a survey conducted among PA students, which received 63 responses, 32.3% of students said they used generative AI at least 1-2 times a week, with 12.9% saying they used it at least once a day. Out of the students that stated they used it, ChatGPT was by far the most popular. ChatGPT is similar to other generative AIs in its water usage, and the training process uses roughly the same amount of carbon dioxide as five average American cars over their lifetimes.
If we want to work towards a more sustainable future, we have to cut back our usage of generative AI. Spreading awareness is a huge way to do this. In the same survey issued to PA students, 55.7% of students said they were not aware that there were environmental issues concerning generative AI. But, after being educated about the environmental issues of generative AI, 52.5% of students said they were less likely to use it. This shows the good that simply spreading awareness can bring.
So, when you go about your day after reading this, I would urge you to be more mindful of how you use generative AI, and inform others of the negative effects that using it can have. It doesn’t take much to make a big difference.
