Artificial Intelligence (AI) technologies have undoubtedly revolutionized various aspects of our lives. From voice assistants to self-driving cars, AI has become an integral part of modern society. However, the environmental implications of AI technologies are a growing concern that cannot be ignored. The growing demand for computing power not only leads to increased freshwater consumption, but also exacerbates air pollution. This problem has caught the attention of researchers at the University of California, Riverside, who argue that large technology companies are not paying adequate attention to fairly distributing these environmental impacts.
Impact on freshwater resources
One of the major environmental impacts of AI is the excessive consumption of fresh water to cool huge data centers. These data centers, which house the powerful computing systems needed to process AI data, require huge amounts of water for cooling. It is shockingly estimated that large data centers can consume 1 to 5 million gallons of water every day. This poses a serious problem, especially in water-stressed regions such as Arizona, where several tech companies are planning to set up data centers. Plans to consume massive amounts of water in a water-stressed desert region raise serious concerns about sustainability and responsible resource use.
Addressing the problem: Geographic distribution of load
In their research paper, “Toward Environmentally Just AI through Geographic Load Sharing,” the UCR team suggests possible solutions to the environmental inequality caused by AI technologies. They argue that technology companies such as Google and Microsoft have the flexibility to distribute their computing power and computational load evenly across the globe. By redirecting computing loads from water scarce regions to water abundant areas, these companies can minimize environmental impact and promote sustainable practices.
The role of electricity consumption
In addition to freshwater consumption, AI technologies also contribute to air pollution through the massive amount of electricity required to run data centers. Many of them are powered by coal-fired power plants that emit not only carbon dioxide but also other harmful substances such as particulate matter and nitrogen oxides. These pollutants have detrimental effects on human health, including increasing the risk of cancer, cardiovascular disease and shortening life expectancy. Populations living near power plants bear the brunt of environmental and health impacts, further exacerbating environmental inequalities.
The need for equitable distribution
The findings of the UCR study highlight the urgent need for large technology companies to address the issue of equitable distribution of environmental impacts caused by AI technologies. International organizations such as UNESCO and the OECD also highlight the importance of mitigating the environmental inequalities of AI. It is critical for technology companies to prioritize sustainability and responsible resource management in their operations.
Conclusion: A call to action
As AI technologies evolve and integrate into our daily lives, the hidden environmental costs associated with them must be recognized and addressed. The equitable distribution of computing power and computational load proposed by UCR researchers is a step toward minimizing the environmental impact of AI technologies. Technology companies must take responsibility and prioritize sustainable practices to ensure a more environmentally just future.
