Samsun AI water footprint calculated by Ondokuz Mayıs University

Professor Yusuf Demir of Ondokuz Mayıs University has calculated the environmental impact of artificial intelligence applications in Samsun, Turkey, specifically focusing on the water and carbon footprint created through energy consumption.

In Samsun, the annual AI water footprint per capita is estimated to range between 785 and 2,714 liters. Demir noted that while calculating the water footprint of AI is not yet a widespread practice, it is theoretically possible by considering indirect factors such as energy consumption and the water used during energy production.

Key factors influencing this footprint include the energy required for large language models and deep learning networks, the cooling systems of data centers, and the types of water resources used (blue, green, and grey water). The impact varies by city depending on the energy type and the specific methods used for electricity production and server cooling.