Artificial intelligence (AI) is rapidly becoming part of our daily lives, with the global AI market expected to reach a value of $250 billion by the end of 2025, according to Statista.
A Large Language Model (LLM) is the most common form of AI and is seen in popular products such as OpenAI’s ChatGPT, Google’s Gemini and Microsoft’s CoPilot.
These LLMs utilize large amounts of computer power to process requests from their millions of users, leading companies to build large data centers all over the country and world.
As the amount of energy usage increases with the rise of AI, critics say the environmental cost is more than most realize.
Northern Virginia is a hub for data centers, with more than 51 million square feet of data center space total, requiring enough electricity to power more than 800,000 homes, according to WTOP News.
Susan Hartley, environmental science teacher, said the large amount of electricity being sent to these data centers can disrupt the flow of electricity in the region they’re located in.
“AI takes much more energy and computer power which pulls much more energy from the surrounding grid,” Hartley said.
With the U.S. removing itself from the Paris Climate Agreement, an international treaty to reduce climate change adopted by 196 countries, some worry the dangers of climate change due to increased energy use will not be taken seriously.
“If we have a society and a government that doesn’t believe in the dangers of climate change, there will be issues,” Hartley said.
OpenAI spends about $3 billion per year to train their ChatGPT models and get them ready for release, according to their own statements. This money is mostly spent on energy costs for data centers as well as research and development.
DeepSeek, a Chinese AI startup, claims they only spent $6 million training their new model released earlier this year, which they say is on par with the most advanced ChatGPT model available.
These factors have raised concerns from the public that the U.S. market is spending and using more than necessary on energy, computer chips, and data center space, which ultimately lead to a sizable tech stock market crash on January 27, according to the Guardian.
Kevin Koch, AP Environmental Science teacher, said energy use is not the only concern when discussing AI, but that e-waste is another major factor. E-waste, or electronic waste, refers to any electrical or electronic device that is thrown away and is particularly dangerous due to toxic chemicals that naturally leach from the metals inside when buried.
“E-waste can contain rare earth metals like lithium, nickel or cobalt that will definitely contribute to harming the environment,” Koch said.
When e-waste is disposed of in an unsafe manner, it can release up to 1,000 different chemical substances into the environment, including harmful chemicals to humans such as lead and mercury, according to the World Health Organization.
Koch said he doesn’t believe the U.S.’ current energy infrastructure can handle the load of the new data centers being built, but said there are ways to ramp up production in sustainable ways.
“If we start to use more renewable methods or even convert older coal burning plants into more efficient and less polluting plants, then the environmental cost of AI will definitely go down,” Koch said.
Drew Hansen, senior, said he uses AI every day to help him with things he doesn’t know the answer to, like a search engine. He said schools should be more receptive to AI and teach more about its positive and negative effects, including its effects on the environment.
“AI is obviously going to be used a lot in the future, so they should teach us more about its negative effects so we can use it in a better way,” Hansen said.
