Whether or not you are conscious of it every time you send a question to an AI chatbot water use, there is a significant amount of water used to attempt to give you an answer. But just how much water is consumed at every inquiry? The answer is contingent upon who is counting.
Why Data centers Require Water
Data centers utilize water to aid in the cooling of the servers, both directly with the water used at cooling towers at the electricity stations that produce steam. In 2021, Google’s data centers in the US alone are reported to have consumed around 12.7 billion liters of fresh water.
What the Statistics Indicate
Preliminary calculations: Research conducted by the University of California at Riverside finds that approximately 0.5 liters are consumed for every 10-50 inquiries made. This number is often misinterpreted as the water consumed per inquiry.
Google: Using information published by Google, the average Gemini query accounts for 0.26ml of water per inquiry. This is equivalent to approximately five drops.
OpenAI: According to Sam Altman, the average amount of water consumed in ChatGPT is equal to 1/15th of a teaspoon of water per inquiry. This statement was made by the company itself.
Why the Estimates Vary
The water consumption of a data center is determined by location, cooling technique, and whether or not the water from the power plant is included. The data from the organizations and scientists is hardly comparable.
What Can Be Done – AI chatbot water use
Scientists believe that AI training should be planned for cooler hours, when water evaporates less. Some companies are experimenting with air-cooling. Water use is not required to be reported in the UK yet.
Stay informed on nature’s biggest global events.
How Harmful Is Plastic Pollution?
Explore its impact on ecosystems.
How Do Communities Fight Drought?
Meet the volunteers building their own dams in Hungary.
Can Renewables Deliver a Greener Future?
