In today's world, it is hard to go a day without encountering artificial intelligence (AI). Whether it is an article about a new use, glitch, software platform advertisement, or social media meme, AI is everywhere—even spell-check on this very article.  

Artificial intelligence has a wide range of applications across various fields, including education, healthcare, customer service, weather forecasting, crop monitoring, energy optimization, cybersecurity, financial trend analysis, voice assistance, autonomous vehicles, and even writing (unfortunately).

AI has undeniably driven incredible advances and will continue to fuel progress in ways we cannot yet imagine. However, there are downsides to AI that are often overlooked. In this article, I will focus on the sustainability implications of these drawbacks.

In the past, computers used individual hard drives to store, access, and process data on your device. Today, artificial intelligence utilizes cloud computing, which allows data to be stored and processed in remote data centers. These data centers provide faster processing, increased data storage capacity, real-time data access, and support for complex algorithms.

Although cloud computing and storage have enabled advancements in the world of technology, they also carry significant environmental consequences. AI can enhance computing power beyond what an average phone or computer can handle. However, this increased capability requires an increase in energy consumption. Estimates show that a single AI query can consume up to 25 times more energy than a standard Google search. This surge in energy use results in a surge in emissions.

The United States' energy grid is already operating at maximum capacity. Without significant infrastructure investment and the ongoing rise in AI usage, we may face power shortages. To combat this electricity shortage, the Trump administration has promoted the use of coal. The U.S. Energy Information Administration reports that coal emits 50% more CO2 than natural gas (EIA report). Not to mention, coal mining is damaging to natural landscapes and has multiple pollution factors.

Additionally, the high processing and energy consumption levels cause servers to overheat, which is often solved through water-intensive cooling systems. This method effectively reduces temperatures, but at the cost of the overuse of another natural resource. For example, Google’s data centers, which support Gmail, Google Drive, and YouTube, use 200 million gallons of water annually. This is equivalent to 303 Olympic swimming pools used for cooling only one of these companies.

While there is no need to fear AI, it is important to understand when and where it should be used. Questions such as the score of a game, the inventor of the toaster, or who appears on the latest reality TV show doesn't require AI to answer. ChatGPT is not the new Google. Using AI effectively means reserving it for situations where traditional resources might fall short. AI should be used only for tasks that are impossible without it, newly enabled by it, or would otherwise require an overwhelming amount of manual effort.

This ensures we're not automating just for convenience but truly unlocking value that couldn't exist before. By focusing AI use on these areas, we can drive innovation responsibly and avoid unnecessary environmental damage.

Shapiro is excited to implement a new project for Earth Day to continue advocating for sustainable data use and management. This year, we aim to reduce our company's data footprint, improve data organization, and help colleagues incorporate these practices into their personal lives as well. We hope to track the related emissions reductions and continue with sustainable data practices going forward. Keep an eye out for our findings!