July 21, 2024
AI's Growing Energy Thirst: Water Consumption in Data Centers Raises Environmental Concerns
AI

AI’s Growing Energy Thirst: Water Consumption in Data Centers Raises Environmental Concerns

More water is also required for cooling the servers that respond to your commands due to artificial intelligence’s increased energy requirements. Additionally, the thirst is only growing.

AI systems like ChatGPT, which use servers frequently housed in expansive data centers that can consume huge quantities of electricity and water, require extensive processing and, consequently, electrical power, just like any computationally intensive activity. The two major players in the space, Microsoft and Google, claim they are reducing the environmental impact of AI as part of a larger industry initiative to use resources more wisely. But experts keep bringing up the drawbacks of unregulated development in the tech industry.

According to a recent article by the Associated Press, ChatGPT and other AI systems like it force data centers to use 500 milliliters of water every time a user asks between 5 and 50 queries or prompts. That’s not a little amount of water, which has experts and environmentalists sounding the alarm. The virtual assistant has become the fastest-growing tech site viewed on Earth since the invention of the internet.

The paper states that ChatGPT’s GPT-4 models were trained on data centers close to corn fields west of Des Moines, Iowa, where water from the Raccoon and Des Moines rivers is used. “It (GPT-4) was made by these extraordinary engineers in California, but it was really made in Iowa,” Microsoft’s president, Brad Smith, reportedly stated, according to the AP.

Since then, the AI revolution’s power-hungry gear has needed more water to run, both to power it up and to cool it down. Official data indicate that Microsoft’s water consumption increased by 34% between 2021 and 2022, probably as a result of the demands of computing connected to AI. Its biggest rival, Google, claimed a 20% increase in water usage within that time.

To put this in context, the two corporations’ combined increase amounts to an extra 1.7 billion gallons of water.

Even though OpenAI has kept the locations of its data centers a secret, the influence could be felt locally. Axios reported in April 2022 that residential prices at Des Moines Water Works had climbed by over 80% since 2007, and were now $5.19 per 1,000 gallons for municipal residents.

This, according to some experts, illustrates the unexpected implications of fast-developing AI technology, which necessitates increasing computer capacity.

“Most people are not aware of the resource usage underlying ChatGPT,” researcher Shaolei Ren told AP. “If you’re not aware of the resource usage, then there’s no way that we can help conserve the resources.”

Other emerging technologies have huge thirsts as well. Even though it has nothing to do with AI, cryptocurrency mining offers a useful analogy. According to a 2021 Cambridge University study, the energy used annually for Bitcoin mining is similar to all of Argentina’s electrical consumption.

This topic has generated debate, but it has also been crucial to the growth of the entire crypto business. During the infamous “crypto winter,” when Bitcoin’s value crashed from over $65,000 to less than $15,000, Tesla discontinued its support for the cryptocurrency out of worries about its effect on the environment. Since that time, the topic of bitcoin mining has been discussed in politics across the world.

Tech firms do note that they are making efforts to lessen their environmental implications. For instance, Google claims that its data centers use 60 percent less energy than the industry standard.

Microsoft stated that it is “investigating how to encourage transparency and efficiency improvements at leased sites” and plans to switch to carbon-free energy entirely by 2030. In its sustainability report for 2022, Microsoft stated that it is “focused on getting our own house in order” and fulfilling its promises to 2030, which include becoming a carbon-negative, water-positive, and zero-waste firm by the end of that year.

Similar to this, Google’s environmental report for 2023 emphasized the company’s objective of attaining net-zero carbon emissions spanning its operations and value chain by 2030, necessitating a swift decrease in the estimated 10.2 million tons of CO2 generated in 2022. Google also pointed out that traditionally, when the requirement for AI computing grows, the energy required grows more gradually than predictions suggest due to efficiency measures.

The environmental effects of AI’s insatiable appetite for data and power are still a concern. Tech companies must diligently explore efficiency improvements, renewable energy, alternate cooling methods, and resource usage transparency as AI capabilities advance. It is possible to sustainably quench AI’s thirst with appropriate supervision.

Image: Wallpapers.com

Disclosure Statement: Miami Crypto does not take any external funding, or support to bring crypto news to the readers. We do not have any conflicts of interest while writing news stories on Miami Crypto.

Related posts

OpenAI CEO Sam Altman’s Worldcoin Project Secures Additional $100M Funding for AI-Related Problems 

Kevin Wilson

Google’s Gemini AI Gets Affordable: A Boost for Developers

Harper Hall

Salesforce AI CEO: AI’s Evolution Paralleling Early Internet

Harper Hall

This website uses cookies to improve your experience. We'll assume you're ok with this, but you can opt-out if you wish. Accept Read More

Please enter CoinGecko Free Api Key to get this plugin works.