The use of artificial intelligence (AI) has long been touted as a solution to reduce humanity’s environmental footprint. However, a recent study by the Yale School of Environment has shed light on the darker side of AI technology, particularly in terms of its energy consumption and environmental impact.
According to the study, the environmental footprint of AI is large and growing, with concerns raised about its excessive water and electricity consumption. Jeremy Tamanini, founder of Dual Citizen, warned that if left unregulated, AI’s carbon footprint could have dire consequences on sustainability and emission reduction targets.
One of the most pressing environmental concerns highlighted in the study is the massive water consumption by AI systems. For example, it was found that 10 to 50 responses from ChatGPT-3 can use up to half a litre of water. Tech giants like Google and Microsoft are using millions of litres of fresh water to cool their data centres, leading to a strain on freshwater resources.
Furthermore, the study pointed out the lack of data on AI’s overall environmental impact, including direct carbon emissions and electricity consumption. Without proper regulations and standards, tech companies have been either reporting arbitrary data or withholding information on their AI’s environmental footprint.
Efforts are being made to develop sustainable AI practices, with proposals for more transparency and access to information on AI’s energy impact. The International Electrotechnical Commission and the International Organisation for Standardization are set to release international standards for sustainability in AI, covering areas such as energy and water consumption, carbon footprint, and waste management.
While the future of AI technology remains uncertain in terms of its environmental impact, there is a growing recognition of the need for more sustainable practices and regulations to mitigate its negative effects on the environment.