The Hidden Environmental Impact of Relying on AI Technologies

Article by Yi Xuan

ChatGPT: Convenience at the Cost of Sustainability

For many people today, particularly the youth, ChatGPT seems to be an indispensable part of our everyday life. We often use ChatGPT for education, work, or even creative use, making our progress much more efficient than before without putting in the same amount of effort. However, with ChatGPT optimising so many tasks for us, it leads us to think

whether such a technology is more of a step towards sustainable innovation or a step back instead?

I argue that while artificial intelligence (AI) like ChatGPT reflects significant progress in

technological innovation and optimisation, its widespread use today, often without awareness of its high energy demands, suggests it does more environmental harm than it contributes to sustainability.

While ChatGPT was well-intentioned to improve efficiency and optimize tasks, its prevalent use today has only made people more reliant on it rather than using it consciously. Many now turn to ChatGPT even for minor tasks, reflecting their growing dependency. Prior to ChatGPT, most of us would search for answers online via search engines or posing questions to others. ChatGPT now uses 10 times the amount of energy compared to a Google Search Function (Rw, 2024), showing how it takes a huge toll on our energy and water resources (Figure 1).

Figure 1: Comparison of energy use between Google Search and ChatGPT. (Source: Kanoppi)

As part of a broader group of AI technologies, ChatGPT relies on energy-intensive data centres to function. Its efficiency in generating rapid responses actually masks an unseen environmental cost. Like many AI applications today, its convenience encourages habitual, unchecked use, creating a digital dependency that exacerbates its already-substantial energy footprint.

This tension between convenience and environmental cost is not unique to ChatGPT, but reflects a larger trend in the AI sector. While AI is often lauded for enhancing efficiency and optimizing electricity use, its widespread and unregulated consumption has produced paradoxical outcomes. Most everyday users are unaware of AI’s environmental impact, and hence have little incentive to reduce their usage. According to the International Energy Agency (IEA), electricity consumption by data centres will more than double by 2030, driven by artificial intelligence (AI) applications that will create new challenges for energy security and CO2 emission goals (The Straits Times, 2025). This surge in demand is mainly due to the energy-intensive cooling systems required to maintain the data centres that support AI model use.

Some might argue that AI can help mitigate environmental impact by enabling smarter resource management. However, while AI can improve efficiency, excessive or careless consumption risks negating those gains. This should prompt us to reconsider the net environmental benefit of the use of AI, and how the environmental burden might outweigh its benefits due to our overuse today.

So, the next time before you ask ChatGPT a question to help you with your assignment, think about this:

Is it really necessary? We can always fall back to our traditional search engines, and use ChatGPT with more intention and consciousness to move towards sustainability. Rather than abandoning AI, we should instead develop a culture of mindful digital consumption, transforming ChatGPT from an environmental burden into a tool that aligns with sustainable innovation.

References

Rw. (2024, November 4). How much energy do Google Search and ChatGPT use? RW Digital - Vancouver Digital Marketing Agency. https://www.rwdigital.ca/blog/how-much-energy-do-google-search-and-chatgpt-use/

The Straits Times. (2025, April 10). AI surge to double data centre electricity demand by 2030: IEA. https://www.straitstimes.com/world/europe/ai-surge-to-double-data-centre-electricity-demand-by-2030-iea

Towler, L. (2025, August 17). Search Engines vs AI: energy consumption compared. Kanoppi. https://kanoppi.co/search-engines-vs-ai-energy-consumption-compared/