Dear Eartha, I use Chat GPT a lot at work, but I hear that data centers use a lot of energy. What’s the environmental impact of AI?
If you have been paying attention to the news the last few years, chances are you have probably heard about Artificial Intelligence (AI). While AI gets a lot of media coverage, the environmental impact of AI isn’t always mentioned. With such a big technology shift on the horizon, let’s dive into what the popularity in AI might mean for the environment.
What is Artificial Intelligence?
The term AI gets thrown around a lot, but many people don’t actually know what it means. Artificial intelligence is a type of technology that allows computers to perform advanced tasks, such as ones normally done by human reasoning. AI can be used for all sorts of projects, from helping you write your grocery list to supporting NASA in diagnosing problems with their satellites. That’s pretty cool!
AI’s Environmental Cost
Something we all might take for granted is that everything we access in the digital realm exists somewhere physically as well. For example, consider the cloud. It might seem like a virtual storage space, but your data actually exists somewhere. These places are called data centers. Like most technology, AI relies on data centers to function. And it turns out, data centers are extremely resource intensive.
Electronics
To operate data centers, you need powerful computers and microchips, all of which require precious metals such as platinum, gold, and silver. But mining these materials does not come without a cost. Mining can destroy important ecosystems, create water pollution, produce toxic waste and release carbon emissions. And as AI gets more popular and the demand for these resources goes up, the mining industry is expected to grow substantially, increasing AI’s environmental impact.
Energy
Perhaps the biggest criticism of AI is the intense energy usage required to train and run its programs. In most cases, this energy still comes from fossil fuels, which generate a large carbon footprint. By 2026, the International Energy Agency predicts that AI will account for four percent of the world’s total energy usage – comparable to the annual energy consumption of Japan.
What’s the carbon pollution impact of all that energy? A study done by the University of Massachusetts discovered that it takes “626,000 pounds of carbon dioxide, or the equivalent of around 300 round-trip flights between New York and San Francisco” to train a singular AI program. That is a lot of carbon emissions.
Water
Since AI is such an energy intensive technology, the data centers require advanced cooling practices to ensure their computers do not overheat. This is most often done by using water to cool the electronic systems. Some of the largest data centers, like ones used by Google, can use upwards of 550,000 gallons of water a day. This becomes problematic when data centers are built in areas like the western US, where water is already a scarce resource.
Is it all doom and gloom?
So, is AI the final straw for the climate? Not necessarily. Some experts are very hopeful that AI could help us solve our climate problems. Just like how NASA can use AI to help with their space missions, AI can help climate scientists better understand climate change. It can help them collect important data, model potential natural disasters and synthesize solutions to that to help stop the climate crisis. It is another tool in the toolbox for activists and scientists to use to help address climate change.
A great example of this is using AI to manage the energy efficiency of the electrical grid. With more renewable energy capacity added to grids nationwide, and increasing demand for electricity as more cars and homes go electric, traditional grid management techniques are falling short. Researchers at University of Virigina have shown that with the use of AI, the electrical grid can better handle variations in demand and optimize energy usage. This AI solution will help ensure the reliability of renewable energy going forward.
AI holds immense potential—both as a powerful tool in the fight against climate change and as a technology with significant environmental costs. The path AI takes will depend on how companies and policymakers choose to move forward. By prioritizing renewable energy and innovations that minimize resource use, we can steer AI toward becoming a solution rather than a setback in our climate journey.
Ask Eartha Steward is written by the staff at High Country Conservation Center, a nonprofit dedicated to waste reduction and resource conservation. Submit questions to Eartha at info@highcountryconservation.org.
