thad edits

Author: willitcode

content/impacts/_index.md

@@ -2,11 +2,11 @@
 title = 'The Impacts'
 +++
 
-As big tech companies across the world turn their focus to the power of AI, the number of resource-intensive data centers needed to support the growing industry only increases. At the moment, data centers account for 1-2% of the overall global power consumption, a number that is predicted to double by 2030 to around 2-4% of the total power demand.[^1] As the power demand for data centers grows however, so do the emissions associated with the generation of the power. Because it is such a massive demand, companies constructing data centers often choose areas with cheap energy, not renewable energy sources, meaning that they rely on coal or oil burning power plants which have massive carbon footprints for their power.[^2] New data centers can often delay the closing of these carbon-producing power plants, or even prompt the construction of new ones to provide the necessary electricity, pumping out more carbon into the atmosphere, continuing to warm the planet and provoke extreme weather events. 
+As big tech companies across the world turn their focus to the power of AI, the number of resource-intensive data centers needed to support the growing industry only increases.[^1] Because data centers require so much electricity, companies constructing data centers often choose areas with cheap, instead of renewable, energy sources meaning that they rely on coal or oil burning power plants which have massive carbon footprints.[^2] New data centers can often delay the closing of these carbon-producing power plants, or even prompt the construction of new ones to provide the necessary electricity, pumping out more carbon into the atmosphere, continuing to warm the planet and provoke extreme weather events. 
 
-With the increase in focus on renewable energy and carbon emissions, one of the biggest problems with artificial intelligence slips under the radar, the water required to cool the servers in the data centers. Throughout the cooling process, up to 9 liters of water per kilowatt of energy used are evaporated, rendered useless to the cooling process and inaccessible for those in need of drinking water. This cycle of high consumption and low recycling creates a deficit that is usually addressed by simply pulling in more water from other sources, draining the surroundings of a much needed resource. Although big companies such as Google, Microsoft, and Meta say that they aim to replenish more water than they consume by 2030, there are doubts if that is an achievable goal given that there is simply not enough water to last until then.[^3] Compounding with the incessant need for more water, data centers are often built in locations with cheap real estate and energy, such as South America and Sub-Saharan Africa, without any thought to the pre-existing water stress of those areas, which leads to a worsening of the problem as water is drawn away from those in desperate need and redirected into cooling the massive data centers. Although there are advances in technology and efficiency in regards to data centers, global water consumption needed for AI is still projected to reach between 4.2 and 6.6 billion cubic meters by 2027.[^4]
+With the increase in focus on renewable energy and carbon emissions, one of the biggest problems with artificial intelligence slips under the radar: the water required to cool the servers in the data centers. A cycle of high consumption and low recycling creates a deficit that is usually addressed by simply pulling in more water from other sources, draining the surroundings of a much-needed resource.[^3] Data centers are often built in locations with cheap real estate and energy, leading to a worsening of the area’s water stress as water is redirected into cooling the massive data centers.[^4] 
 
-While AI is a source of hope for many in regards to potential solutions for the climate crisis, we must not ignore its negative impacts on the environment. The power and water required to keep the data centers, and by extension AI, functional, creates numerous problems across the globe. The prioritization of cheap land and power over sustainable practices means that data centers rely heavily on non-renewable sources of energy that produce massive amounts of greenhouse gasses, and place incredible amounts of stress on the surrounding watersheds. Although generative AI practices may create solutions to environmental crises in the future, we must not ignore the current dangers and practices inherent in the current practices.
+Overall, the environmental impact presented by data centers across the world contribute to preexisting geographic inequalities. Because of the choices made in the construction of data centers, the negative impacts such as the pollution produced in the process of producing the necessary power and the contamination of the water used in the cooling process are often shouldered by less fortunate countries and people. Freshwater makes up only 3% of our water supply, 2.5% of which is locked in frozen tundras and glaciers. With only 0.5% of our world’s water supply available for human consumption,[^4] a water guzzling industry such as Artificial Intelligence is simply not sustainable. While carbon and greenhouse gasses are the focus of much criticism and vigilance in a newer, more environmentally conscious society, it is vital that we acknowledge the massive impacts of excessive water use and turn our eyes towards its effects as well.
 
 [^1]: Lohr, Steve. “Will A.I. Ruin the Planet or Save the Planet?” *The New York Times*, The New York Times, 26 Aug. 2024, www.nytimes.com/2024/08/26/climate/ai-planet-climate-change.html?searchResultPosition=9.
 [^2]: “Google blames AI as its emissions grow instead of heading to net zero” *AlJazeera*, AlJazeera, 2 Jul. 2024, https://www.aljazeera.com/economy/2024/7/2/google-blames-ai-as-its-emissions-grow-instead-of-heading-to-net-zero