Artificial intelligence (AI) not only
raises fears for jobs and human rights, but it also risks having
a huge impact on the environment.
Experts estimate that the demand for AI will increase
groundwater or surface water withdrawal to between 4.2 and 6.6
billion cubic metres by 2027.
This is roughly half the amount consumed by the United Kingdom
each year.
Moreover, the boom in generative AI could increase the demand
for energy and water even more.
The issue lies in the so-called large language models that are
at the heart of the new technology and are capable of processing
large amounts of textual, numerical and other data.
To function, these models require huge amounts of computing
power and the use of server farms that employ chilled water,
some of which can be reused, to cool equipment by absorbing heat
from the air.
The Financial Times reports that in 2022 the growing use of data
centres led to an increase in water consumption of 34% for
Microsoft, 22% for Google and 3% for Meta.
Overall, according to the International Energy Agency, data
centres, cryptocurrencies and AI accounted for almost two
percent of global electricity demand in 2022.
This figure could double by 2026 to reach almost the electricity
consumption of Japan.
Sam Altman, the father of ChatGPT, raised the alarm at the World
Economic Forum's annual meeting in Davos, saying an energy
breakthrough is needed in order to power future AI.
The OpenAI CEO also said, however, that this breakthrough should
be powered by climate-friendly energy sources such as nuclear
fusion or the cheaper solar power.
ALL RIGHTS RESERVED © Copyright ANSA
