AI is becoming an integral part of everyday life; from the workplace and education, to the tools we use and the questions we ask daily. One of AI's newest applications relates to water conservation. While many regions globally face water scarcity, South Africa ranks among the top 30 driest countries due to low rainfall and worsening factors such as climate change, rapid population growth and poor water management. Given that approximately 79% of South Africa's land is utilised for agriculture, AI could potentially play a pivotal role in securing the country's water future.
Recent policy developments indicate that the South African government is increasingly aware of the strain AI infrastructure can place on national resources. In May 2024, the Department of Communications and Digital Technologies published a National Policy on Cloud and Data, urging data centres to reduce their reliance on the national electricity and water grids. While not yet legally binding, the policy highlights a broader commitment to sustainability and calls for operators to implement alternative and environmentally friendly solutions. As South Africa continues to develop its national AI policy framework, more targeted regulation around AI's environmental footprint, particularly water usage, should follow.
South Africa does not yet have specific legislation regulating AI in the context of water use by data centres. However, existing environmental and water-related regulations still apply. Data centres, including those hosting AI infrastructure, are subject to environmental legislation covering water use licences, waste management, and energy consumption. As AI operations grow and the demand for water increases, especially for cooling large-scale computing facilities, these regulations play a vital role in balancing technological advancement with environmental sustainability.
Conversely, the integration of AI may assist with water conservation when linked to services such as leak detection, irrigation optimisation, demand forecasting, and public communication on outages. Investment in the water sector is increasingly focused on technological advancements, particularly AI.
In theory, AI has significant potential to address water scarcity, including optimising water distribution networks, improving water quality and treatment, and even potentially developing drought-resistant crops. This could involve predictive maintenance of infrastructure, real-time monitoring of water usage, and even developing new technologies for desalination and wastewater treatment.
Other than the costs associated with advanced technology, one major drawback of AI infrastructure is the heavy water consumption required to run the data centres. Fresh water is a limited resource, and many regions already struggle with shortages. As AI expands, its demand for water increases, putting even more pressure on local supplies. For example, in 2023, Google's Council Bluffs, Iowa, data centre consumed 980 million gallons of water. In 2022, Microsoft's water consumption surged by 34%, reaching nearly 1.7 billion gallons, largely due to its growing AI operations. Looking at ChatGPT, the carbon dioxide emissions per year amount to 8.4 tons. Additionally, a single user's conversation of 20-50 questions uses around 500ml of water making the total water consumption substantial when accounting for its billions of users. This level of consumption can create competition between tech companies and local communities, especially in regions already dealing with droughts. Greater AI-related water use reduces availability for households, agriculture, and ecosystems.
However, there are ongoing efforts by tech companies to mitigate these negative impacts. Companies are starting to take notice of their water usage and are implementing strategies to reduce their water footprint. For instance, Google has committed to replenishing 120% of the water it consumes by 2030, and Microsoft is investing in water-positive technologies, and it too aims to achieve this by 2030. AI is also used to optimise water use in data centre cooling systems, significantly reducing consumption.
With South Africa still developing the infrastructure necessary to sustain AI data centres, we must also take precautions to not lose sight of the environmental toll that powers an AI-driven future. While AI offers remarkable solutions to water scarcity, its infrastructure risks becoming part of the problem if left unchecked. South Africa stands at a crossroads: either lead the way in sustainable AI development or allow innovation to drain one of our most precious resources. The challenge is undeniable, but so is the opportunity.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
