It’s abundantly clear that not everyone is a proponent of artificial intelligence. The technology’s popularity with ordinary people remains tenuous, to say the least, though it remains to be seen whether that popularity waxes or wanes over the coming years when clear use cases for the technology become more apparent.
For now, however, one of the most prevalent ways to criticize AI companies is based on their use of energy, particularly the large data centers used to power the technology. We don’t intend to discuss that here, as it has, indeed, been reported in great detail elsewhere. Instead, we wanted to ask the question of whether AI is forcing all of us to be more reflective on our energy use.
There is growing evidence that we are becoming more conscious as a society of energy use. Of course, not all of this is down to AI. Rising energy bills – a theme of the 2020s globally – are clearly influential in making us think about energy use. Yet, perhaps a better way to frame it is that AI can make us consider pointless energy use.
Digital energy consumption is a nuanced debate
Of course, concepts like digital carbon footprints predate consumer-focused AI. Moreover, it is often an area of nuance and cannot be simplified to say something like using a computer is bad for the environment. Consider, for example, the carbon footprint of getting into your car and driving to a casino for the afternoon compared to playing slot games online at DraftKings. The former is clearly much worse for the environment.
Yet, what we mean by pointless uses of energy has become a hot-button topic of late. During those viral AI image generation crazes of last year, for instance, it emerged during the Studio Ghibli viral image creation – where users turned themselves into Japanese anime characters – that each image generation used the energy equivalent of fully charging a smartphone. 100s of millions of these images were generated, so you can appreciate the cumulative effect.
<iframe width=”560″ height=”315″ src=”https://www.youtube.com/embed/b0C56yqIkbk?si=taYc03CKaJ2Pc1H_” title=”YouTube video player” frameborder=”0″ allow=”accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share” referrerpolicy=”strict-origin-when-cross-origin” allowfullscreen></iframe>
While digital energy use has often been a talking point in the internet era, the use of water by data centers is something that feels relatively new with AI. Those image generations mentioned above? They can use anything from 5 to 50 liters of freshwater per image. Sea water can’t be used because it’s highly corrosive and leaves behind so much debris. It’s become a huge point of contention as more and more huge data centers pop up to support the AI boom.
AI energy demands will grow in the coming years
What is also acutely clear is that the need for energy to fuel the AI boom is only going to get greater. Billions of dollars, perhaps trillions, are being invested in data centers and other infrastructure to meet demand over the coming years. It’s not just energy use, either, as those data centers can also produce harmful wastes. So this topic is perhaps going to become more prevalent in the coming years. Auxiliary pressures on energy sources, such as the conflict in the Middle East, don’t help either.
So, should we feel guilty and stop using AI? Not necessarily. The UN and other organizations suggest a more pragmatic approach. They recognize the power of AI, claiming that, indeed, the technology might help solve many of the environmental issues we face globally. Yet, they also counsel that we should all become more conscientious about how and when we use the technology.
Indeed, we could even follow the advice of AI bosses like Sam Altman, who have said that we should refrain from saying please and thank you to ChatGPT. Altman pointed out that it costs the company a fortune for those additional prompts, but for the rest of us, we should be aware that each time we hit “enter,” it consumes a lot of energy.