It is the dawn of a new era. With the inception of Chat GPT in 2022, the rise of artificial intelligence (AI) has grown rapidly. For college students, this has provided a new tool for essay assistance and quick access to information.
However, as we increasingly rely on these tools, there may be a hidden cost to their convenience. By saving time, we might be sacrificing our intellectual growth, and the environment could be paying the price.
There has been an increasing concern that the use of AI in schools is unethical. Students may use technology for projects or essays and take credit for the content generated by AI without acknowledging its use. As a result, instructors have begun reevaluating how they create assignments, with some now asking students to watch a video and connect its content to course concepts, rather than simply assigning a written essay. Additionally, there has been an increased demand for tools that detect the usage of AI. However, while stopping the use of AI in academia has become increasingly complex, there is also an adverse side effect that students may not be aware of.
By offloading cognitive tasks to computers, people stop exercising these mental skills. This is not only true for AI; in the 1990s, many people could memorize telephone numbers, but with the increased use of cellphones, this skill has dwindled. Research has shown that pawning off higher-order critical thinking tasks to generative AI impacts students’ ability to articulate thoughts and could negatively impact their critical thinking during oral presentations. This has raised concerns that an over-reliance on generative AI may hinder the development of skills such as problem-solving and critical thinking. As such, AI should be utilized only for routine and mundane tasks, rather than for complex, critical thinking. See A Primer on Generative Artificial Intelligence
In addition to the cognitive impacts, there are also hidden costs of AI related to the high energy required to perform its complex, intensive algorithms. For example, according to Faisal Kalota from the Center for Information and Communication Sciences, training the GPT-3 model consumed “1.287 gigawatt hours and 700,000 liters of clean freshwater…[enough] to power 120 U.S. homes for one year.” This process generated approximately 552 tons of carbon emissions, equivalent to the yearly emissions of 120 U.S. cars. Even the seemingly simple task of asking ChatGPT a question can produce 100 times more carbon emissions than conducting a single Google search. One AI data center consumes as much energy as heating 50,000 homes for a year. See Tackling AI’s Climate Change Problem
Even the single image included in this article, created by Google’s Gemini AI image generator, is said to have consumed around 0.01 to 0.29 kilowatt-hours of energy. That is the equivalent of running a refrigerator for up to half an hour. See The Hidden Cost of AI Images: How Generating One Could Power Your Fridge for Hours
Looking at the broader picture, the United States has an ambitious goal of reducing greenhouse gas emissions by 50% from 2005 levels by 2030, as supported by the Paris Agreement. AI has made this goal increasingly difficult to reach, as AI already accounts for 0.6% of global carbon emissions, and this figure is expected to rise unless urgent action is taken. According to the International Energy Agency, to meet the United States’ climate change goals, AI’s energy footprint must be halved. This rapid increase in energy demand presents numerous environmental challenges, and companies such as Amazon, Google, IBM, Microsoft, and OpenAI are at the forefront.
The decisions these organizations make to address AI’s energy demand and its impact on climate change will have far-reaching implications for future generations, and the world is witnessing these decisions unfold in real time. One significant move occurred in September 2024, when Microsoft announced that the Three Mile Island nuclear power plant would reopen to meet the growing energy needs of its artificial intelligence programs. Constellation Energy, the plant’s owner, plans to rename the facility Crane Clean Energy Center, with an expected reopening in 2028 . See Three Mile Island nuclear plant will reopen to power Microsoft data centers
While nuclear energy has long faced public skepticism due to concerns over safety and potential disasters, such as Chernobyl and Fukushima, recent research into nuclear energy has shown that it is a clean energy source, and continued technological advancements can further reduce its environmental impact. See The Role of Nuclear Energy in Reducing Greenhouse Gas (GHG) Emissions and Energy Security: A Systematic Review
Nuclear energy is a highly efficient source of power, capable of producing large amounts of electricity while emitting minimal greenhouse gases. Although renewable energy sources like wind and solar are integral to reducing emissions, they cannot provide the reliable, consistent energy needed by AI. AI’s energy demand is continuous, without regard for time of day or weather conditions, making intermittent renewable energy sources less viable. See Nuclear Energy
Unlike renewables, Microsoft has identified nuclear power as a stable and predictable energy source, vital for meeting its continuous energy demands—essential for mitigating the effects of climate change.
As the next generation of innovators, we can help drive the shift towards a more sustainable future. By being mindful of how we engage with AI and supporting cleaner energy options, like nuclear power, we can pave the way for a tech-driven world that fosters intellectual growth without sacrificing the health of our planet.