Can AI help us fight the ongoing global warming?

Could the explosion of technologies of recent years be the answer to the climate crisis? This is what several studies and reports on artificial intelligences and on their usefulness in the fight against the deregulation of weather. A named study Tackling Climate Change with Machine Learning and published in 2019 had in particular been talked about because it explored the possibilities offered by artificial intelligences, more and more present in our daily lives. Properly controlled, these could help fight against global warming.

AI, a tool in the fight against the climate crisis

Today, artificial intelligence is present everywhere: in algorithms for proposing content, on platforms such as TikTok, YouTube or Netflix, with the increasing automation of certain models of cars like the Tesla or through voice assistants like Siri, Alexa or Google Home. The installation of AI in modern society allows a greater comfort of life, but can also be useful for solving problems. ecological issues. The study Tackling Climate Change with Machine Learning, directed by David Rolnick, professor at McGill University in Montreal, focuses the use of AI on several areas concerned by ecological transition: transport,energy, the construction cities or industry.

In some regions of the world, artificial intelligence is already being used for the ecological transition. The Chinese juggernaut Tencent announced in 2020 to embark on the construction of a ultramodern city, Net City, ecological and car-free, governed by technology and artificial intelligence. The Chinese utopia, which is estimated to cost $ 1.2 billion, is still being worked out and is based on the success of automated and autonomous ports spreading along the Chinese coast. In parallel, nearly 500 smart cities are currently created in China, and should allow the superpower to “green” its metropolises, often classified as most polluting in the world.

At the same time, energy renovation could also benefit from the efficiency of artificial intelligence. The residential sector emits 20% of greenhouse gas at the national level, many buildings require a considerable energy input and firms propose to increase the energy performance of dilapidated installations thanks to new technologies. AIs are now being created to map and acquire data on the different areas to be renovate because no longer corresponding to current energy standards, as proposed by the company NamR.

At the industry level, the use of artificial intelligence could make it possible to improve production conditions: find the least polluting materials, carry out shorter and less polluting supply chains. Areas, such asagriculture, could also benefit from the advantages related to AI, by implementing a technique called “precision farming”. The University of Sydney thus designed the robot Rippa, powered by solar energy and equipped with a hyperspectral camera, the latter allowing to target areas in which spreading from pesticides need to be done. Autonomous, Rippa is programmed to perform tasks such as weeding or general maintenance of agricultural plots.

The tertiary sector, bringing together 75% of French workers, is not left out: AI could eventually provide viable management of resource management, such as the purification of waste where the recycling from waste. The Toulouse-based company Ffly4u, for example, offers a box attached to the site buckets to collect data on the container of these buckets and thus offer a tracking garbage. It would then be possible for building companies or waste reception centers to find out more about the quantities of waste to be recycled or to offer new storage solutions.

To set up these autonomous systems, it is necessary to resort to Machine Learning (or machine learning in French), a method allowing computers use data on a subject in order to continuously learn and improve using statistical and mathematical calculations. Improved AI is also synonymous with better simulations concerning climatological changes and helps to anticipate the impact of the climate crisis on continents and populations. Artificial intelligences therefore represent real progress, but also have an energy cost that could weigh against them.

A strong ecological toll?

While artificial intelligence therefore makes it possible to find answers to certain major issues, it does create other problems: the production of terminals and other IT equipment such as data centers that leave a strong imprint carbone all the more important since these devices are created and assembled in countries of Central and Southeast Asia, where energy consumption can be extremely polluting.

On the other hand, the use of digital equipment and the resulting electricity consumption only represent a marginal share of emissions of CO₂. In addition, with the rise of artificial intelligence and automation, data centers should experience an optimization of their energy performance. According to a study commissioned by the Senate in 2019, pollution linked to data centers could well increase by 86% by 2040 due to the increase in the consumption of digital services by French households, but the study does not include the gains obtained by optimizing the performance of data centers, nor those of the energy mix French which gives pride of place nuclear and which is therefore very low in carbon, offsetting the pollution induced by the use of artificial intelligences. The future will tell if these new initiatives will help civilization to emerge from the climate crisis.

All the articles of the special day: