A recent research from the University of California, Riverside, published in The Washington Post, highlights the hidden environmental consequences of employing AI technology such as GPT-4. While many people believe AI just affects computer power, the study identifies another significant resource consumption: water. AI systems require a lot of water to cool the computers that process data, such creating text using GPT-4. This procedure also requires a significant amount of power, which puts further strain on resources, particularly in places where AI data centers are situated.
The study discovered that the quantity of water required to create 100 words with GPT-4 varies by region, based on proximity to the data center and local energy prices. For example, in Texas, 235 milliliters (a little less than a cup) of water is required to generate a 100-word text output, but in Washington state, the same output requires around 1,408 milliliters—about three 16.9-ounce bottles of water.
Although these sums may appear insignificant at first, they build up rapidly when considering the millions of requests made to GPT-4 every day. When AI is employed on a regular basis, the environmental impact becomes apparent. Furthermore, issues extend beyond water use. Data centers’ high power consumption have an impact on adjacent towns, resulting in increased utility costs.
The study offered an example in which Meta (previously Facebook) utilized 22 million liters of water to train its LLaMA-3 model. This quantity is equivalent to the water necessary to cultivate nearly 4,400 pounds of rice, or what 164 Americans would consume in a year.
Electricity usage is another issue. If only 10% of the U.S. working population (about 17 million individuals) utilized GPT-4 once a week, yearly power use would exceed 121,000 megawatt hours. This is similar to the amount of power consumed by each residence in Washington, D.C. For twenty days. Given the predicted large use of GPT-4 by enterprises and individuals, the analysis finds that this power use estimate is surprisingly low.
Despite these mounting worries, the main AI companies, including OpenAI, Meta, Google, and Microsoft, have been largely unclear about how they intend to handle these challenges. While they have publicly committed to reducing environmental problems, it is unclear how they will do so. For example, Microsoft spokesperson Craig Cincotta noted plans to develop water-free cooling systems, but provided no details on how or when this would happen.
Overall, although the benefits of AI grow, the environmental costs are becoming more important. These hidden expenses raise concerns about sustainability as tech corporations advance AI advances.
This article is based on research shared by The Washington Post and written by Christopher Harper. You can check out the full article here.
Hi, I'm Voss Xolani, and I'm passionate about all things AI. With many years of experience in the tech industry, I specialize in explaining the functionality and benefits of AI-powered software for both businesses and individual users. My content explores the latest AI tools, offering practical insights on how they can streamline workflows, boost productivity, and drive innovation. I also review new software solutions to help readers understand their features and applications. Beyond that, I stay up-to-date with AI trends and experiment with emerging technologies to provide the most relevant information.