1.7 billion liters of water: this was necessary to develop Microsoft’s AI
Artificial intelligence does not come without costs, especially impacts on water consumption. In their latest environmental reports, both Microsoft and Google revealed significant spikes in water consumption due to the use of artificial intelligence (AI) technology and this was picked up by an AP article.
Almost 6.4 billion liters of water
Microsoft reported a 34% increase from 2021 to 2022, equivalent to nearly 1.7 billion gallons (6.4 billion liters) of water, while the latter reported a 20% increase over the same time period. All this due to the need to cool the electronic computers they operate
“It's fair to say that most of the growth is due to artificial intelligence,” including “its heavy investments in generative artificial intelligence and partnership with OpenAI,” Shaolei Ren, a scientist at the University of California, told AP , Riverside, which has been on a mission to evaluate the environmental impact of AI.
An incredibly expensive training course
OpenAI's GPT-3 training consumed 700,000 liters of water, "enough to produce 370 BMWs." The research also estimated that ChatGPT, which followed GPT-3, should have consumed the equivalent of a bottle of water. 500 milliliter water to carry out a discussion with a user of approximately 25-50 questions. A person who answers 25-50 questions doesn't consume all this water…
Large-scale cooling solutions are needed in data centers, where AI calculations are performed, to prevent servers and other devices that power the technology from overheating. HVAC (heating, ventilation, and air conditioning) systems and cooling towers use a significant amount of water.
Additionally, AI calculations can be energy-intensive. In some places, producing the electricity needed for these operations can involve significant water consumption, especially when using power generation technologies such as thermoelectric plants, which rely heavily on water for cooling.
AP reached out to Microsoft, which responded with a statement saying the company is "working on how to make large systems more efficient, both in training and application."
To achieve its sustainability goals of becoming carbon negative, water positive and waste-free by 2030, the company said it will continue to analyze its emissions, increase the use of clean energy to power data centers, purchase renewable energy and to invest in other eco-friendly initiatives. However, the problem of water resources remains.
Meanwhile, Open AI told the news outlet that the company “recognizes that training large models can be energy and water intensive” and is actively working to improve efficiency.
Solutions explored
One solution being explored is to reduce the energy requirements of AI calculations. These efforts can make AI more energy efficient by building better algorithms and hardware, which can indirectly reduce water use. Additionally, placing data centers in areas with access to renewable energy sources could reduce AI-related water use.
There is also a more creative option. In 2018, Microsoft sank one of its data centers off the coast of Orkney and studied its performance for two years. The center was found to have operated more efficiently during that time period.
Part of the reason may have been that the water surrounding it kept it cool, Microsoft said at the time of its recovery in 2020 that the process had environmental benefits. “The recovery began the final phase of a years-long effort that demonstrated that the concept of underwater data centers is feasible, as well as practical from logistical, environmental and economic standpoints,” Microsoft's John Roach wrote in a blog.
So the solution could be to put computer centers in cold regions. Who knows, maybe one day they might think of using these centers for co-generation too..
Thanks to our Telegram channel you can stay updated on the publication of new Economic Scenarios articles.
The article 1.7 billion liters of water: this was needed to develop Microsoft's AI comes from Economic Scenarios .
This is a machine translation of a post published on Scenari Economici at the URL https://scenarieconomici.it/17-miliardi-di-litri-di-acqua-questo-e-stato-necessario-per-sviluppare-la-ai-di-microsoft/ on Mon, 11 Sep 2023 07:00:40 +0000.