The evolution and advancement of technology have two sides to them. On one hand, we witness how they transform various sectors and accelerate growth and productivity. On the other hand, these advancements often come at the cost of our natural resources. This is the story of OpenAI, the creators of the popular AI chatbot ChatGPT, which has dominated the charts since its debut. This technology consumed a significant amount of water in Iowa to satisfy its data requirements. The artificial intelligence technology behind ChatGPT, a large language model capable of generating human-quality text, was developed in Iowa with substantial water usage.
Microsoft, which owns a majority stake in OpenAI, in its latest environmental report revealed that its global water consumption jumped 34% from 2021 to 2022 – a huge increase from previous years. According to OpenAI (via Associated Press), the company used a cluster of data centers in West Des Moines, Iowa to train the language model. The data centers used about 11.5 million gallons of water in July 2022, the month before OpenAI says it completed the training of GPT-4.
That amount of water is equivalent to about 6% of all the water used in the West Des Moines Water Works district, which also supplies drinking water to the city’s residents.
The water was used to cool the powerful supercomputers that were used to train ChatGPT. The supercomputers generated a lot of heat as they processed the massive amount of data needed to train the language model.
ALSO READ l ChatGPT hype over? OpenAI’s popular chatbot sees monthly website visits decline three months in a row
The West Des Moines Water Works in 2022 in a document said that it and the city government “will only consider future data center projects” from Microsoft if those projects can “demonstrate and implement technology to significantly reduce peak water usage from the current levels” to support the water supply for residential and other commercial needs. Microsoft is said to be working on its response to this feedback.
The massive consumption of water to train AI models is a growing concern. As AI technology becomes more sophisticated, it requires more computing power, which in turn requires more water. This calls for steps like using renewable energy sources to power the supercomputers and conducting more research into the environmental impact of AI technology. As AI becomes more powerful, it is important to find ways to reduce its environmental impact.
With inputs from Associated Press
Follow FE Tech Bytes on Twitter, Instagram, LinkedIn, Facebook.
