Experts and Researchers Debate Environmental Impact of Artificial Intelligence

The environmental impact of artificial intelligence is a subject of increasing scrutiny, with experts highlighting both its potential to mitigate climate risks and its capacity to deplete natural resources. At a science workshop held at the Madanapalle Institute of Technology and Science, experts stated that AI could improve flood prediction, water management, and the monitoring of climate and resources. Professor D. Nagesh Kumar of the Indian Institute of Science noted that combining AI with satellite imagery and real time sensor data could enable smart water resource management.

Conversely, other reports emphasize the ecological costs associated with the infrastructure required to run AI. Data centers housing the computers necessary for AI generate significant heat, requiring cooling systems that often rely on fresh water. According to a study from the University of California, Riverside, training a large model like GPT 3 in a typical data center could directly evaporate approximately 700,000 liters of fresh water. Total global data center water consumption was estimated at around 560 billion liters by 2023, with a significant portion coming from indirect use via the electricity supply chain.

Researchers suggest that the primary concern is not the global loss of water, but the depletion of fresh water in specific locations, particularly in areas already facing water stress. While some experts view AI as a tool for environmental protection, others argue that its real impact remains unknown and urgent measurement is required.