Despite what it seems like, Artificial Intelligence, or AI, is not new, with a history spanning over 70 years — long used by scientists to process and analyse data to come up with insights or predictions.
The Massachusetts Institute of Technology (MIT) defines AI as computers’ ability to perform human-like functions such as learning and problem-solving.
Virtual assistants like Siri, Alexa, and recommendations on apps such as YouTube are just a few modern-day examples.
However, what seemed like science fiction in the 1950s has now evolved into today’s Generative AI (GenAI), which is seeing a massive boom in users with the accessibility of apps such as ChatGPT, Gemini, and DeepSeek.
Known for its ability to generate text, images, music, and code using data patterns, GenAI — while useful — is highly controversial for its negative environmental impact.
Today, AI is almost inescapable, with mainstream apps such as Google and Facebook integrating it into the lives of even the unacquainted, potentially causing more harm to the environment.
Source: Tech Xplore
According to MIT News, one ChatGPT query uses about five times more electricity compared to a simple web search.
This is alarming since generating electricity from fossil fuels, the most widely used means, involves greenhouse gas emissions that make the planet hotter and releases pollutants into the air and water, said Dr Yuan Yao, associate professor of industrial ecology and sustainable systems at Yale.
Source: Airam Dato-on on Canva
AI also requires an exorbitant amount of water to cool data centres that house equipment essential to its function.
An International Energy Agency (IEA) report on energy and AI noted that an average 100-megawatt data centre in the US can consume as much electricity as 100,000 households annually and about two million litres of water per day.
Source: baranozdemir on Canva
In Singapore, 500ml of water is used for every 34 prompts sent to ChatGPT-3, the Organization for Economic Co-operation and Development (OECD) revealed.
The environmental costs of everyday AI use are only part of the picture.
Making new data centres, producing Graphics Processing Units (GPUs), and training AI models to meet the increasing demand also contribute to the pollution of our planet, as these require the use of toxic chemicals and thousands of megawatt hours of electricity.
Some regions are bearing the brunt of AI’s negative effects on the environment more than others, something that is not getting as much attention as it should.
Most AI data centres in both the US and Asia are built and continue to pop up in arid areas with scarce water supplies.
Southeast Asia also faces this problem as Microsoft and Google have committed billions of dollars in investments in data centres in the region.
Source: AirTrunk via iTnews Asia
Data centres in Malaysia, for example, have put a strain on their resources and caused the demand for water to exceed its capacity, according to Fair Planet.
As such, there is a risk that treated water meant for public consumption may be used instead.
Additionally, data centres in torrid areas not only compromise locals’ access to clean water, but also affect agriculture and energy production.
Source: Rachmattt on Canva
Despite recent initiatives to make AI sustainable, some areas are also overlooked.
For instance, Google’s data centre in Finland operated on 97% carbon-free energy in 2022, but its data centres in Asia were only 4% to 18% carbon-free, Harvard Business Review reported.
AI also has positive impacts on the environment — but whether it is worth its risks is the question.
According to research published in the scientific journal Hygiene and Environmental Health Advances, AI can accurately detect pollution, monitor air and water quality, and forecast natural disasters.
Compared to sending out researchers to study these, AI is also faster and can easily cover large geographic areas and handle copious amounts of data.
Source: Barcelona Supercomputing Center
Tech companies have also begun experimenting with other methods to reduce their reliance on clean water to cool down their data centres.
Amazon Web Services uses treated sewage water during certain times of the year and is hoping to work with local utility companies to gain access to more recycled water.
Microsoft and other tech companies are also planning to build closed data centres that prevent the evaporation of water and allow it to circulate between servers and chillers without having to be refilled.
However, an investor in one of OpenAI’s sites admits that this system would require even more energy than evaporative methods, Bloomberg reported.
Needless to say, the journey towards sustainable AI is long, and it is uncertain whether our planet can wait.
AI is certainly not without merits.
For instance, a team from MIT recently created an AI model that can predict if a patient is predisposed to develop breast cancer based on a mammogram.
This proves that there is so much potential in this technology that can positively impact us.
Source: MIT News
However, majority of users are also using AI to answer simple questions and generate cartoon images of themselves, which are contributing to environmental decline.
So perhaps it is best to leave this technology to professionals.
As Dr Noman Bashir, Computing and Climate Impact Fellow at MIT Climate and Sustainability Consortium, puts it:
The ease-of-use of generative AI interfaces and the lack of information about the environmental impacts of my actions means that, as a user, I don’t have much incentive to cut back on my use of generative AI.
As such, AI companies must be held accountable with stricter regulations requiring them to reveal how their processes can affect the planet, and the measures they are taking to make their environmental impact a net positive.
Also read: MS Explains: How common is breast cancer in S’pore & are you at risk of getting it?
Have news you must share? Get in touch with us via email at news@mustsharenews.com.
Featured image adapted from Matheus Bertelli on Canva.