Produced by: Tarun Mishra
All AI requests, from generating images to answering queries, are processed in data centres that use significant computing power and resources.
These data centres consume vast amounts of energy, often sourced from fossil fuels, nuclear energy, and other natural resources, to power supercomputers.
Research by Sasha Luccioni and her team at Hugging Face revealed that training the GPT-3 model for ChatGPT produced 500 metric tonnes of carbon dioxide, equivalent to over a million miles driven by gasoline-powered cars.
Contrary to popular belief, digital requests are processed at physical data centres, which require substantial energy for storage and processing.
Data centres generate significant heat and require cooling, often using fresh water and traditional air conditioning systems, increasing the strain on local resources.
Google’s data centres in Dalles, Oregon, were found to consume a quarter of the town’s water supply, leading to local water shortages.
AI technology aids in climate solutions, such as predicting weather patterns and mapping flood-affected areas, which are crucial for disaster management and agricultural planning. While AI is essential for various applications, sustainable and responsible use is necessary to minimize its environmental footprint and ensure long-term benefits.
The environmental cost of technological advancements, similar to the steam engine and cryptocurrency, underscores the need for responsible AI use to avoid exacerbating environmental distress.
To mitigate AI’s environmental impact, transparency from tech companies about their carbon footprint and responsible usage practices are essential. Global efforts should focus on treating AI as a valuable resource rather than exploiting it for trivial applications.