How Artificial Intelligence Does Not Destroy The Environment

Artificial Intelligence can be thought of in the same way as cloud computing. What this basically claims is that if you view them through an environmentalist’s point of view, an impending and violent source of carbon emissions emerges, with the possibility to destroy all clean air and replace the strengths of the world for good.

While this viewpoint appears quite excessive and over-the-top, many environmentalists do share this exact viewpoint towards AI being used in climate change. They claim that AI has done more harm than good to our climate, and view the constant improvement in technology as directly proportional to an environment in ruins and shambles. However, while considering recent improvements in AI’s application towards the environment, and the words of established researchers who have studied this field, this claim appears biased, outdated and unreasonable.

The previous goal of AI with the environment was solely focused on how AI could make current carbon-emission-generation-pastimes smarter. Letting AI lead these specific power loads instead of a variety would lead to greater efficiency, less electricity usage by humans, and without a doubt, minimize the chances of the environmental destruction the environmentalists believed in. Yet, while this occurred, the main concept that was unaccounted for was how much electricity would AI truly use. While it’s true humans would use less electricity, it was difficult to discover AI’s perspective in electricity, whether that be even larger, also less or the same as in previous experiments.

But now, fortunately, with Gary Dickerson’s work towards his company Applied Materials, a major semiconductor producer and thus an automatic expert in the demand for circuitry required to handle the transfer of data, an answer is observed. The electricity AI would use would roughly be 10% of the world’s electricity use by 2025, unless super-efficient semiconductors are innovated within these next couple of years.

His forecast should heavily be considered by environmentalists who believed that AI would take over a much greater amount of electricity in the future than just 10%. In fact, many other CEO’s of notable AI or tech related companies have viewpoints even less severe than Dickerson. These include the usage of electricity remaining flat, at around 1-2%, due to Google’s innovation to use AI technology to cool down its data centers, reducing a heavy power demand in different google labs around the world by 40%.


A cooled-down Google data center

A cooled-down Google data center

At the same time, many other data centers, among the most hottest in Utah and Nevada, lack these innovations to cool down AI and keep the data center rooms running longer. It’s important that AI technology is at a cool temperature so that it manages longer and more efficiently. But one thing is for sure. Despite this, each of the positive realities, from Dickerson’s perspective to the other CEO’s, it is very unlikely that artificial intelligence will cause a havoc in society. More and more progress must be completed to continue reducing the usage of electricity by artificial intelligence, naturally, but there are few chances, if any, for any whopping increase.