AI’s insatiable thirst for energy threatens to dynamite Google’s climate goal of net zero carbon emissions by 2030.
Ecology report Published by Google This week, instead of falling, greenhouse gas emissions are expected to increase by 13% in 2023 compared to the previous year. Compared to the reference year of 2019, the increase reaches an impressive 48%.
AI applications run in large data centers and require more processing power -hence energy– about the online services we know so far.
Three-quarters of the electricity Google uses comes from renewable sources, while the rest is produced in natural gas or coal-fired power plants that emit carbon dioxide and methane.
Data centers also require high voltage lines and large amounts of water to cool computers. They are also noisy.
Because of the explosive spread of new technology, some experts worry that artificial intelligence threatens the transition to clean energy, a necessary step to combat climate change.
In fact, the International Energy Agency (IEA) predicts energy consumption in data centers could be doubled Until 2026.
“Achieving the 2030 goal is an extremely ambitious mission,” Kate Brandt, Director of Sustainability at Google, admitted to the Associated Press.
“We have to deal with great uncertainties, such as the uncertainty about the future of artificial intelligence’s impact on the environment.”
According to the report, most of Google’s data centers in Europe and America run on RES and do not aggravate the climate. But the same is not true for data centers in the Middle East, Asia and Australia, many of which run on fossil fuels.
Of course, the problem does not only apply to Google, but also to other technology companies. In May, Microsoft reported that its emissions increased by 29% this year compared to the 2020 reference year.