Inspenet, November 25, 2023.
A study by Shaolei Ren, a researcher at the University of California, reveals that large models of artificial intelligence, such as ChatGPT, which go through an extensive training process that involves thousands of computers operating constantly, contribute to the increase in the consumption of water used for cooling these systems.
The magnitude of the problem intensifies when considering the total impact, since it would be necessary to take into account the water used in the generation of electricity to power the servers or in the hardware manufacturing processes. Companies such as Amazon, Microsoft and Meta (parent of Facebook) control the majority of hyperdata centers in the world, raising additional concerns about the sustainability of the use of water resources.
The preferred method of cooling these data centers is by using water , which poses additional environmental challenges. The communities that host these centers, often located in water-stressed areas, face massive evaporation of water during the cooling process. This liquid, contaminated with minerals and salt, is discharged into wastewater plants, limiting its potential reuse for human consumption.
Despite calls to decrease the use of digital tools as a means to mitigate this problem, specialists argue that data centers are vital in contemporary society. With the industry continuing its commitment to the integration of artificial intelligence in various products and services, concern about the sustainable use of water in these operations is presented as a constant challenge.
The controversy related to the demand for artificial intelligence raises essential questions about the environmental responsibility of technology companies and underlines the need to comprehensively address this problem to ensure sustainable development in the digital age.
Don’t miss any of our posts and follow us on social media!