The abbreviation “AI” needs no introduction in 2025. Artificial intelligence is now as ubiquitous as smartphones, and has been implemented into almost every aspect of technological development. The powers of AI span from helping write a brief email to detecting early stages of breast cancer, saving the lives of millions. Open-source AI has become a tool for students, professionals, and everyday consumers. It’s commonplace to see advertisements promoting AI’s ability to write emails, take notes, or create art. Critics have raised concerns over the morals of this new technology. However, artificial intelligence comes with more ethical costs. It actively contributes to climate change.
To understand the impacts of AI on the climate, it is important to understand how artificial intelligence works. Early home computers came with large, bulky, data storage servers that would overheat when overworked. Eventually, these data servers were outsourced to data centers, massive complexes of computers that store and compute data. Artificial intelligence derives all of its computations from the servers in these centers. The massive amount of energy required causes the servers to heat up significantly. To prevent the servers from crashing, data centers pump massive amounts of water through the servers to cool them down.
In the area around Phoenix, Arizona, there are more than 58 data centers, which can equate to more than 170 million gallons of drinking water used per day for cooling data centers, which puts a strain on the already fragile water supply. The placement of data centers in fragile ecosystems has raised concerns about the entanglement of business in local politics. Google faced heavy criticism from residents of Mesa, Arizona when plans to build a new data center there were unveiled. The kicker? Google would only have to pay $6.08 per 1,000 gallons of water compared to the resident price of $10.80 per 1,000 gallons. As desert regions become drier and hotter from changing climate patterns, pressure on water availability grows. The preferential treatment large corporations experience over residents in these high-risk regions raises concerns about water rights.
Data centers do not just guzzle up viable drinking water, but also use extraordinary amounts of electricity. According to the International Energy Agency, a request made through ChatGPT, an AI-based virtual assistant, consumes 10 times the electricity of a Google Search. This vast amount of energy often entails increased emissions of fossil fuels, since most of the power grid in the United States comes from coal or oil power plants. If data centers do not cool their servers using water cooling systems, they rely on HVAC units and centralized air. This system of cooling also comes with high energy costs, growing AI’s emission rates.
Most companies do not disclose information about their emissions reports, including ChatGPT owner OpenAI. This can make it difficult for consumers to know if companies are hitting their emissions targets. Google claims to have hit net emissions since 2007. However, they have only hit this target from purchasing carbon offsets. Carbon offsets are tradable rights corporations buy to fund projects that lower carbon dioxide levels in the atmosphere instead of taking actions to reduce their own emissions. A recent report by Google detailed a 48% increase in emissions since 2019. Similarly, Microsoft saw a 24% increase in emissions in 2020, which they state is due to the construction and operation of data centers to support their growing AI workloads.
Tech companies have not just used fossil fuels to help power their growing AI models. The same AI models are actively helping the big energy industry more efficiently drill and mine non-renewable resources. Microsoft entered an agreement with the oil company ExxonMobil to use Microsoft’s cloud-computing platform Azure to help optimize mining operations to increase oil production by 50,000 barrels per day this year. Microsoft is helping add more crude oil to the market, and their emissions to the atmosphere. AI could be used in deforestation practices, monocropping in farms, and increasing efficiency in clothing productions, all of which harm the environment.
AI is not a tool created with malicious intent, and its applications are not just limited to fossil fuel optimization or questionable generative art. Artificial intelligence can be an important tool in helping to solve issues exasperated by the growing climate crisis. Researchers have been able to use AI to track methane emissions from melting permafrost and manage food scarcity in at-risk populations. When contacting Scientific American, a Microsoft spokesperson stated that “technology has an important role to play in helping the industry decarbonize, and this work must move forward in a principled manner—balancing the energy needs and industry practices of today while inventing and deploying those of tomorrow.”
The biggest issue with artificial intelligence is the lack of transparency. Big tech companies have neglected to disclose their environmental impacts in public reports, blinding consumers to the impacts of artificial intelligence use. They dupe consumers by pushing AI products that are designed to increase convenience in daily life. AI is a powerful tool that has the potential to benefit humanity and the environment. However, the commercialization and commoditization of AI as a tool to make our lives “easier” does more environmental harm than good. Using ChatGPT to write a basic email may seem like the cheap way out to save a few minutes, but thousands of people using AI for simple tasks compile over time. Consumers need to ask themselves: is AI enough of a necessity in my life to justify its negative costs on the environment?
