Did you know that before reading this blog, you wasted almost 1 litre or more of water by searching?
Your every chatGPT search takes 500 ml of fresh water. Imagine that every time you use CHAT-GPT to look something up or ask a question, it uses up millions of liters of water. This might sound strange because we usually think of using the internet as something that only affects the environment a little. But behind the scenes, running and cooling the powerful computers that make CHAT-GPT work uses much water. So let’s delve into how Chat-GPT, a popular language model, utilises a vast amount of water for each search, shedding light on the unseen consequences of our digital interaction. Our lives have changed significantly in the last few years due to the amazing development of artificial intelligence. Nowadays, artificial intelligence is a part of our daily lives, providing us with intelligent virtual assistants, tailored ideas, and answers to a variety of queries.
Our prominent example is how people now use AI-powered tools like chatGPT for information searches, sometimes preferring it over traditional search engines like Google. This success of AI, particularly OpenAI’s ChatGPT, is Largely due to its ability to generate text responses that sound remarkably human-like, using simple and natural language. People think that is the future of AI.
But did you know that ChatGPT “Drinks” around 500 ml of fresh water for every 20 to 50 questions answered?
A 500 ml bottle of water might seem like a little but the problem is created if you consider the total water ChatGPT uses regularly and how many people use it for daily searches.
AI models like ChatGPT are powered by massive data centers housing thousands of high-performance servers.
These servers process immense volumes of data, executing complex calculations to generate responses. This intensive computational activity generates significant heat and it is unbelievable. To maintain this extreme heat the servers need advanced cooling systems to maintain optimal operating temperatures. Water is crucial in these cooling systems, often used in vast quantities to dissipate heat effectively. A study done by the Natural Resources Defense Council (NRDC) estimates that data centers in the US alone use about 1.8 billion gallons of water annually.
Data centers that host AI systems, use a lot of water, particularly for cooling in hot climates. A report from 2022 shows that the tech giant Google’s data centers used more than 21 billion liters of drinking water, a 20% increase over the year before. Making AI hardware, like semiconductors and microchips, also needs a lot of water, and this process can cause water pollution and environmental damage. Even training AI models uses a vast amount of water. Around 500 milliliters of water are used by ChatGPT for every five to fifty prompts. When you take into account how many people use it, this adds up to a lot of water. An ideal ChatGPT user visits nearly six pages in a single visit on the website, while the average number of pages per visit per user is 6.04 minutes.
The Effects of AI on the Environment
Training and using AI models need a lot of computer power, which usually comes from non-renewable energy sources. This leads to high greenhouse gas emissions.AI technology also creates electronic waste. Disposing of this waste can be harmful because it contains dangerous chemicals. Systems use a lot of energy to run, adding to overall energy use and emissions. Based on the reports scientists believe that this will create many problems in the upcoming days.
The environmental impact of AI, especially its water use, is a big concern. Data centers that run AI systems, use a lot of water mainly for cooling. This water usage is similar to other tech industries, with data centers being the main users.
We must adopt eco-friendly practices and better cooling methods to lessen this impact. Making data centers more efficient, using water-saving technologies, and following sustainable practices can greatly reduce AI’s water use.
To Read more blogs like this Click Here
Writer,
Arnab Chakraborty
Intern, Content Writing Department
YSSE