AI models like ChatGPT ‘drink’ an enormous volume of water, finds study

A recent study by the University of Colorado Riverside and University of Texas at Arlington in the United States studied the “water footprint” that large artificial intelligence models like OpenAI’s ChatGPT leave during their training. The water consumed to run ChatGPT which is used by billions of users across the world, is “extremely large,” said the study in a pre-print article titled “Making AI Less Thirsty.”

The study sheds light on the “water footprint” and how these AI models’ consumption of water has “remained under the radar,” said the researchers. Italso distinguishes between “withdrawal” and “consumption,” as the former refers to the physical extraction of water from rivers, lakes, and other sources, as opposed to the latter which relates to the water loss due to evaporation when used in data centres.

ALSO READ |Report sounds alarm over humanity's 'vampiric' overconsumption of its 'lifeblood' water

Add WION as a Preferred Source

According to the researchers at the US-based universities, when we are addressing consumption in this context, “water cannot be recycled,” which is where most of the AI’s water usage is concentrated. The yet-to-be peer-reviewed research, posted as a pre-print in arXiv also estimates a conversation comprising 20-50 questions with the AI chatbot in a single system may “drink” a “500ml bottle of water”.

“While a 500ml bottle of water might not seem too much, the total combined water footprint for inference is still extremely large, considering ChatGPT’s billions of users,” said the researchers.

The water consumed also refers to clean freshwater (to prevent corrosion and bacterial development) which the data centres use for generating electricity as well as for cooling servers to run AI models, which was studied using a framework created by the researchers.

ALSO READ |World Water Day 2023: Parched! The drying future of water

The study claims that while training GPT-3 alone, Microsoft may have consumed a staggering 700,000 litres of water, which is the same amount used to make 370 BMW cars.

Microsoft, which has invested billions of dollars, has also partnered with ChatGPT creator OpenAI. Meanwhile, scientists have also noted Google's Language Model for Dialogue Application (LaMDA) which is used by the giant’s AI chatbot Bard can also consume a “stunning” amount of water in the order of millions of litres.

Therefore, researchers have expressed their concerns over the massive consumption of water by these AI chatbots and called for the companies to “take social responsibility” and address their “water footprint”.

ALSO READ |AI might crack your password in less than a minute. Here's how you can secure your account

According to the study, if the data had been generated in the tech giants’ less energy-efficient data centres in Asia, water use may have increased threefold. The number may also increase with the newly-launched GPT-4 AI system which has a larger model size, finds the study.

“AI models’ water footprint can no longer stay under the radar,” said the researchers, adding that water footprint must be addressed “as a priority as part of the collective efforts to combat global water challenges”.

WATCH WION LIVE HERE

You can now write for wionews.com and be a part of the community. Share your stories and opinions with us here.