ChatGPT has drawn worldwide recognition for its quick and comprehensive answers to a wide range of queries. It is the fastest-growing consumer application ever with over 100 million monthly active users.
However, its success has come at the cost of enormous water consumption, raising environmental concerns.
Here’s what you need to know!
An eye-opening study
A new study called "Making AI Less Thirsty", which was conducted by researchers from the Universities of Colorado Riverside and Texas Arlington, reveals that training AI models such as ChatGPT and Bard use thousands of tonnes of water.
The paper also mentions that "ChatGPT has to 'drink' a 500 ml bottle of water for a simple chat of around 20–50 inquiries and responses. While a 500ml bottle of water might not seem too much, the total combined water footprint for inference is still huge, considering ChatGPT’s billions of users.
Huge amount of Water usage
Researchers estimated water usage shows that Microsoft, a partner of OpenAI, used a staggering 700,000 liters of water just for training GPT-3.
In order to put things into perspective, this amount of water is enough to produce 370 BMW cars or 320 Tesla electric vehicles. It is also comparable to the volume of water required to cool a nuclear reactor.
But, where does the water go?
Data centers that house training data for artificial intelligence are cooled in a variety of methods using water. The two main cooling methods used in data centers — air-cooled and water-cooled systems — both require water.
Air-cooled systems use fans to move air around the servers and other equipment. Water-cooled systems use water to absorb heat from the machinery and transfer it to a chiller or an external cooling tower.
The data centers cannot just use any water. To avoid bacteria and corrosion, it must be pulled from an incredibly clean and fresh water source. Freshwater is also essential for humidity control in the rooms.
Who knew that asking questions and receiving assistance from ChatGPT could be such a “thirsty business”?
The environmental concerns!
But on the brighter side, the water isn't gone forever. Instead, it gets released into the air through cooling towers, although it takes substantial time to return as rain.
Despite this, it's important to recognize the environmental impact of such high water usage, especially as AI becomes more widespread.
Researchers recommended that tech giants must take early action to reduce their water footprint and prevent further harm to the environment.
Comments