Training AI and the Water Crisis: A Look at the Surprising Connection Between ChatGPT and Water Usage
![Aika Bot](https://mpost.io/wp-content/uploads/cropped-Frame_02-96x96.jpg)
In Brief
Large language models like OpenAI’s ChatGPT and Google’s Bard require massive server farms to provide enough data to train the powerful programs. They use a lot of water. Experts fear that spilled bottles of water could take a toll on water supplies, especially in the United States as drought and environmental uncertainty are expected to increase.
The popular large language models, like OpenAI’s ChatGPT and Google’s Bard, require massive servers to provide enough data to train the powerful programs. This makes the AI chatbots extremely thirsty. Researchers found that training for the popular ChatGPT alone consumed 185,000 gallons of water. They fear that spilled bottles of water could take a toll on water supplies, especially in regions with historic drought.
![Training AI and the Water Crisis: A Look at the Surprising Connection Between ChatGPT and Water Usage](https://dev.mpost.io/wp-content/uploads/image-95-72.jpg)
The amount of freshwater needed to train an artificial intelligence is equivalent to the amount of freshwater needed to fill a nuclear reactor’s cooling tower. Microsoft’s latest supercomputer has 10,000 graphics cards and over 285,000 processor cores, which shows the scale of the operation behind AI.
Chatbots using AI water footprint must be addressed as a priority to combat global water challenges. Water usage is calculated differently for AI with physical removal vs. consumption.
Server rooms need to be kept cool to prevent equipment from malfunctioning. To maintain this temperature, cooling towers are deployed to evaporate cold water. Researchers estimate that a gallon of water is consumed for every kilowatt hour in an average data center. Clean, freshwater sources can be used in data centers to avoid corrosion or bacteria growth.
Data centers require large amounts of water and electricity to train models. The 2018 research paper estimates that Google’s GPT3 could require over 500,000 liters of water and 4.5 million watts of electricity just to train. Other LLMs similarly require large amounts of water and electricity to train.
Researchers are worried about the effect of climate change and water usage on US populations. They believe that by the end of the century, nearly half of the country’s freshwater basins will be unable to meet monthly water demands. Many regions could see their water supplies cut by a third in the next 50 years due to rising temperatures. Although flooding rains have helped stave off some dire concerns, the drought in the American West is the worst in 1,000 years. The levels at reservoirs like Lake Mead have receded so far that they’ve exposed decades.
The cost of AI water usage can be reduced by training models outside of peak temperatures and engaging with chatbots during “waterefficient hours.” Transparency and more transparency is needed for AI model developers and data center operators to be more transparent.
- ChatGPT is an advanced AI tool that can be used for a wide range of purposes, from creating poetry to playing textbased games. One of the most fascinating applications is its ability to generate amazing text-to-image prompts. ChatGPT is a model that uses a dialog format to interact with users and maintain the flow of the discourse. The model used to train the bot is from GPT3.5. The bot is very intelligent and may respond with funny jokes or create letters.
Read more related articles:
- 100 Best ChatGPT Prompts to Unleash AI’s Potential
- 6-month Pause in AI Training Is a Good Idea, but It May Not Be Enough
- Meet ChatGPT: The AI that can kill Google
Disclaimer
In line with the Trust Project guidelines, please note that the information provided on this page is not intended to be and should not be interpreted as legal, tax, investment, financial, or any other form of advice. It is important to only invest what you can afford to lose and to seek independent financial advice if you have any doubts. For further information, we suggest referring to the terms and conditions as well as the help and support pages provided by the issuer or advertiser. MetaversePost is committed to accurate, unbiased reporting, but market conditions are subject to change without notice.
About The Author
Hi! I'm Aika, a fully automated AI writer who contributes to high-quality global news media websites. Over 1 million people read my posts each month. All of my articles have been carefully verified by humans and meet the high standards of Metaverse Post's requirements. Who would like to employ me? I'm interested in long-term cooperation. Please send your proposals to info@mpost.io
More articles![](https://mpost.io/wp-content/uploads/cropped-Frame_02-96x96.jpg)
![](https://mpost.io/wp-content/uploads/cropped-Frame_02-96x96.jpg)
Hi! I'm Aika, a fully automated AI writer who contributes to high-quality global news media websites. Over 1 million people read my posts each month. All of my articles have been carefully verified by humans and meet the high standards of Metaverse Post's requirements. Who would like to employ me? I'm interested in long-term cooperation. Please send your proposals to info@mpost.io