7 in 10 Southeast Asia unaware of how much energy AI uses - Singapore News
· The IndependentSINGAPORE: A recent survey conducted by Milieu Insight has shown that most Southeast Asians don’t know how much energy is used up by Artificial Intelligence (AI).
An April 2026 study from Milieu of respondents across Singapore, Malaysia, and Thailand showed that 7 in 10 are unaware of the amount of energy AI really uses. The study also showed that less than a third (29%) correctly know that an AI chatbot uses 10 times more energy than a Google search, with 35% thinking they use about the same.
While a hyperscale data centre uses more than five Olympic swimming pools of water every single day, more than half of the respondents (51%) underestimated this. In Singapore, which had the highest number of participants who knew how much water data centres need, only 51% were aware of this.
Singapore also had the highest trust in government (19%) to act on sustainability, while in Malaysia and Thailand, 12% and 7% respectively say they are “very confident” on the issue.
A majority across the three countries (68%) say that individuals and companies must act to tackle environmental issues, and two in five say they want more accessible green choices for everyday choices, such as affordable eco-friendly products, better public transport, recycling and waste infrastructure that’s easy to use, sustainable food options that are widely available, and energy-efficient products and solutions.
Why is AI so energy-intensive?
A lot of energy goes into operating AI systems as it requires massive computational power to run, train, and maintain. Many systems demand thousands of specialised processors that run continuously in data centres, and generative AI tasks are particularly heavy.
In addition, carbon emissions from power generation, water use for cooling servers, hardware manufacturing and e-waste, and the rapid growth in demand should also be taken into account.
According to one Penn State study, 4.4% of US electricity was consumed by data centres in 2023. By 2028, this figure could triple. This should be a cause for concern, especially since the data centres primarily operate on fossil fuels.
Many scientists have become increasingly concerned, especially climate scientists, sustainability researchers, and energy experts, and there have been calls for monitoring, regulation, transparency, and efficiency improvements.
What individuals can do
While most of the monitoring of AI usage will be done at the government level, individuals who are concerned about the toll AI takes on the planet could practice intentional usage, in the sense of not generating unnecessarily large images or long videos.
People can also urge companies and governments to disclose information about energy and water use, emissions, and the sourcing of electricity, which helps with transparency and accountability. /TISG
Read also: Concerns over Artificial Intelligence in recruitment grow among HR professionals in Singapore
- Advertisement -