ChatGPT consumes a water bottle for every 20 to 50 questions you ask

While ChatGPT has managed to creep into our lives inevitably, there are some interesting facts that you, as a user of this generative AI chatbot, may not be aware of. Recent research found that for every 20 to 50 questions you throw at ChatGPT, it “drinks” a bottle of fresh water. Well, obviously, it does not drink water like you and I do, but it needs that quantity of water to cool the racks of servers as they heat up while ChatGPT tries to answer your questions.

According to a study published by researchers at the University of Colorado Riverside and the University of Texas Arlington in the US, ChatGPT leaves a water footprint, which ultimately impacts the environment. That is because the servers that keep the AI chatbot running need fresh water to run efficiently. In a conversation that includes anywhere between 20 and 50 questions, ChatGPT produces enough heat that needs about 500ml 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, the researchers mentioned in the study.

According to them, it may have taken Microsoft to use whopping 700,000 litres of water just during the training of GPT-3 alone. Next-generation models, such as GPT-4, could consume more water during their training, researchers say, highlighting that there is barely any publicly available data available to give an accurate estimate.

Researchers urged the companies responsible for maintaining AI systems need to “take social responsibility and lead by example” to reduce the water footprint, especially when the global climate crisis is progressing to its next stages. Many companies, including Apple, Google, and Microsoft, have pledged to reduce their carbon footprint as they tread on the journey to become carbon-neutral in the coming years. However, none of these companies highlights the problem associated with high consumption of fresh water in the face of global shortages.

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