IMF: AI's economic benefits will likely outweigh its emissions costs
The International Monetary Fund (IMF) said that the benefits of artificial intelligence would boost global production by 0.5% per year between 2025-2030, and outweigh the rising costs associated with the carbon emissions from the data centres required to run AI models.
The IMF released a report at its annual spring meetings in Washington that noted the fact that these output gains were not shared equally around the globe. It called on governments and businesses to minimize costs to society.
The report stated that despite challenges associated with higher electricity prices and greenhouse gases emissions, AI is likely to be able to offset the costs of additional emissions.
The report "Power Hungry - How AI will Drive Energy Demand" stated that "the social cost of the extra emissions are minor in comparison to the economic gains expected from AI. However, it still contributes to the alarming build-up of emissions."
The adoption of AI will drive a surge in the demand for data processing power that is energy intensive in the coming years. This is happening even though there are still many countries struggling to meet their carbon emission reduction promises.
IMF's report stated that the area dedicated to warehouses filled with servers in Northern Virginia, the place where the largest concentration of data centers is located, was equivalent to eight Empire State Buildings.
The estimated AI-driven global energy needs could triple to 1,500 terawatt hours (TWh) in 2030, which is about the same amount as India's present electricity consumption. It would also be 1.5 times more than the expected demand for electric vehicles during the same time period.
This rise in carbon emissions will be partly determined by whether or not tech companies can deliver on their promises to reduce emissions from data centers through increased use of renewable energy and other methods.
COULD AI lead to energy efficiency gains?
According to the IMF, a strong uptake of AI under current energy policy would result in a global increase of greenhouse gases of 1.2% between 2025 and 3025. It estimated that greener energy policies could limit this increase to 1.3 Gt.
The extra cost was calculated at between $50.7 and $66.3 billion, which is less than the income gain associated with the 0.5% annual increase in global GDP that AI would bring.
Analysts say that the impact of AI on the economy and environment will be heavily dependent on its use. This includes whether AI can result in more efficient energy consumption or sustainable consumption patterns.
Grantham Research Institute on Climate Change and the Environment stated that it could lead to a reduction in carbon dioxide emissions overall if it accelerates advances in low-carbon technology in the power, transportation and food sectors.
Roberta Pierfederici, Grantham Policy Fellow, said: "Market forces alone will not be able to drive AI's application towards climate action."
She said that governments, tech companies, and energy companies all need to play a role in ensuring AI technology is used in an intentional, equitable, and sustainable way. She also cited the importance of R&D funding, and policies, in order to combat inequalities caused by AI advancements. Mark John, London (Writing & reporting) Edited by Ros Russell
(source: Reuters)
