Is it possible to reconcile these technologies and their energy-intensive aspect?
We love, especially in fiction, to depict a world in black and white or on one side the good guys and on the other the bad guys. This is a classification that reassures us, that allows us to believe that good will triumph over evil as in a large majority of works of fiction. This allows us to imagine that we are on the right side. Except we only need to take a step back to realize that it does not work like that. Our world is one of nuances where gray dominates with different shades sometimes darker or paler.
This is evident in all issues, including climate change. While it seems clear that ecologically it is time to act, not everyone agrees on the approaches. Several solutions raise ambivalent feelings.
The possibilities of big data and AI…
For over a decade now, technologies have been able to probe everything quickly and in real-time. We have never had so much data on space, environmental environments, people, etc. This massification of data is seen everywhere, including in the environment. Big data is still rare in rural areas and is not used as much as one might think for monitoring wildlife and plant populations. Nevertheless, it exists and its uses are increasing with satellite applications, among others.
Thus, many believe in the possibilities of this data for companies to reduce their greenhouse gas emissions, to better organize transportation, to better plan urban areas, etc. The manufacturing sector views the use of such information to reduce waste and pollution in companies very favorably. Having access to so much information would make it easier to create examples of the circular economy.
Meanwhile, other companies are aiming to monitor the planet to detect methane leaks or roughly measure carbon capture for plants around the world. Even African countries are benefiting from massive information in the development of their agriculture.
Artificial intelligence, too, fed by data, generates a lot of hope among some in the ecological improvement of the planet. Generative AI, for example, would be able to quickly calculate the effects of a decision by modeling what could happen and thus help in the planning of various policies. We would thus be able to improve the road network for greater fluidity or to perceive oil spills or pollutants in the ground. The goal with data and AI for its supporters is to offer the general public and decision-makers rational and fair information to act accordingly. Except that this has a cost.
… energy-intensive solutions like never before
If the Internet were a country, it would be the third largest consumer of electricity after the United States and China. Data centers that accumulate and participate in big data and artificial intelligence would represent at least 4% of global energy consumption. A value that will only increase given the increasingly frequent uses of the Internet by everyone, including objects. The three major pollution factors are energy consumption, the need to cool electronic equipment and the fact that they are constantly used.
What’s more, the development of data centers was initially done without thinking. The territories allowed them to settle without thinking about the needs, the effects on the surrounding communities, on local energy consumption, etc. Moreover, no longer letting the industry go without putting in place reservations, limitations and others could be a good way to start reducing the ecological footprint of the centers.
Fortunately, some are looking into solutions to reduce the environmental impact of data centers. Already, companies are making or have made a shift to renewable energy and are also improving their design to facilitate server cooling. The goal is to reduce the electricity consumption required to keep IT equipment at a suitable temperature.
There are also solutions within the servers themselves. Many perform unnecessary tasks that technicians could deactivate and the majority of centers are over-equipped, with more equipment in use than necessary. The implementation of more virtual servers would also be interesting since the latter go into standby mode when they are not used and do not consume as much. Currently, they represent half of the servers in a center. This ratio could be increased.
In addition to these changes, we will need to think about our uses. Of course, limiting Internet users’ bandwidth does not really make sense in a world that is constantly becoming digital. However, perhaps we need to think about and inform people about periods of high consumption, encourage the reduction of unnecessary uses of generative artificial intelligence, and offer less polluting digital solutions that favor greener servers. This seems necessary to continue to benefit from the advantages of technology by reducing the ecological footprint.
