Do you know ChatGPT drinks 3.5 million liters or 185000 gallons of fresh water?
Researchers have exposed that the water feasting of data centers mandatory to influence the AI chatbot ChatGPT by billions of users globally is enormously bulky.
The ChatGPT is an AI program with increased consideration for its aptitude to create rapid and wide-ranging responses to various inquiries. With more than 100 million once-a-month active users, it is the wild increasing consumer application in history.
In the meantime, researchers renowned that these accomplishments may have originated from the chatbot’s expenditure consuming enormous amounts of water, around 3.5 million. At the same time, former studies emphasized the carbon tracks of AI models.
Scientists appealed that water used to run them on a big scale has endured under the radar. The new study recommends that a conversation with an AI chatbot in a single system may drink a 500ml bottle of water for 20 to 50 interrogations.
For the study, a new framework was produced to know the quantity of hygienic freshwater expended to bring out the electricity to influence the data center servers and for chilling servers to run AI models.
However, the scientist described it through an example by saying that the working out of GPT 3 alone, Microsoft allegedly consumed 700000 liters (185000 gallons) of water that authority to produce almost 370 BMW cars.
Particularly Google LaMDA is alleged to consume a spectacular quantity of water in the order of millions of liters. The researchers suggested companies use AI models to take social responsibility and address their water footprint in comeback to the world’s water problems.
AI models should also take obligation and lead by example in cooperative efforts to battle the worldwide water shortage challenge by cutting their water track.
A 500ml bottle of water might not look too much; the total combined water footprint for an interface is enormously huge considering ChatGPT’s billion users. Scientists were also implicit that these numbers might intensify by numerous times newly launched GPT-4 AI systems with large model sizes.
Scientists claim ample operative data and the proficiency of water usage throughout the operation of such systems to grow the slide of the water track of AI models.
