Artificial intelligence is revolutionizing and shaping the future of humanity across nearly every industry, including the water sector. It is the primary catalyst for emerging technologies such as big data, robotics and Internet of Things (IoT). In recent months, generative AI has captured widespread attention with tools such as ChatGPT being used on a daily basis by millions of people, solidifying AI’s role as a forefront technological innovator.
Nevertheless, we are currently facing significant challenges regarding AI and its impact on water and energy consumption, as a recent Forbes article highlighted. Let’s look at the data: A single ChatGPT conversation uses around 50 centiliters (cl) of water. It also consumes over half a million kilowatt-hour (kWh) of electricity daily, which is an astonishing amount capable of servicing approximately 200 million requests. Moreover, its daily power consumption is nearly equivalent to that of 180,000 U.S. households, each using about 29 kWh.
Furthermore, AI's water footprint varies considerably depending on its training and hosting location, found a recent OECD policy report. For instance, AI consumes between 1.8 and 12 liters of water for each kWh of energy usage across Microsoft's global data centers, with Ireland and the state of Washington exhibiting the highest and lowest water efficiency, respectively.
This immense water usage and energy consumption is worrying given the rapid expansion of AI product innovations, not only from OpenAI but also from major industry players like Amazon, Anthropic, Cohere, Microsoft, and Nvidia.
According to Cindy Gordon, CEO and Founder of SalesChoice, to meet the demand for advanced AI models, significantly more efficient genAI infrastructures capable of compressing complex models are needed. Additionally, it's imperative to design technology innovations that are both energy-efficient and environmentally friendly. She stresses that water must absolutely be a priority in this endeavor.
She goes on to say that worldwide governments must be aware of these risks and more pressure must be placed on technological giants to act more responsibly and ethically on sustainability matters.
This immense water usage and energy consumption is worrying given the rapid expansion of AI product innovations, not only from OpenAI but also from major industry players like Amazon, Anthropic, Cohere, Microsoft, and Nvidia
In Europe, the EU is set to give final approval to the artificial intelligence law of the 27-nation bloc on Wednesday, paving the way for the world-leading regulations to come into effect later this year. This significant milestone comes five years after the initial proposal of the Artificial Intelligence Act in the European Parliament.
The law will serve as a benchmark for governments worldwide who are struggling to regulate this fast-developing technology.
On Tuesday, Thierry Breton, Commissioner for Internal Market of the European Union, released a statement on behalf of the European Commission highlighting that the Commission is: “addressing the growing concern over the carbon footprint of artificial intelligence (AI) through several initiatives. Key among these is the focus on developing energy-efficient AI technologies supported by Horizon Europe.
“The EU has also set eco-design minimum efficiency requirements for servers and computers that are currently under review and is in the final stages of preparing new rules to monitor the energy performance of data centres to collect and publish data, including also on the energy and water footprint of data centres.
“At the same time, new planning and assessment rules have been adopted with a view to encouraging new data centres to be located where waste heat can be re-used and energy and water needs for cooling be reduced.”
The AI Act is anticipated to officially pass into law by May or June, following the completion of remaining formalities, including approval from EU member nations. Implementation will occur gradually, with countries mandated to prohibit restricted AI systems six months after the regulations are enacted.
Regarding enforcement, each EU member state will establish its own AI oversight body, allowing citizens to lodge complaints in cases of suspected rule violations. Additionally, Brussels will establish an AI Office responsible for enforcing and overseeing the law at a general level.
