“Potential artificial intelligence Helping address the climate emergency is non-negotiable. “In fact, I would go so far as to say that we cannot combat climate change without the help of artificial intelligence.” This is how Nuria Oliver, co-founder and director of the ELLIS Foundation and doctor from the Massachusetts Institute of Technology (MIT) in the USA, expresses herself. Although it is still a technology in its development phase, there seems to be no limit to the possibilities it opens up to achieve the goals set in the 2030 Agenda. And they touch countless areas, from transportation to agriculture to smart building construction to hurricane forecasting and fire management It is destructive in its very beginning stages. A true crossover revolution.
Those who will benefit most from its emergence will be public administrations. Having better weather forecasts, more information about the origins of greenhouse gases released into the atmosphere, and precise tools to mitigate the environmental impacts of highly polluting sectors will help governments fine-tune policies aimed at containing the climate emergency.
The biggest advantage of artificial intelligence is the ability to process very large amounts of data. Past, present and future. And more importantly, build predictive models with them. This way, patterns and trends can be identified to make more precise and informed decisions. «We need digital technologies and large amounts of data that must be collected and analyzed quickly. Artificial intelligence or machine learning algorithms can help us manage this complexity and achieve zero emissions», emphasizes Hendrik Zimmermann, digitalization and sustainability researcher at the Germanwatch NGO.
Special applications
A study published in the journal MIT Technology Review Based on research of lists by David Rolnick (University of Pennsylvania) A dozen specific applications where AI-generated solutions can help combat climate change. Many of them are related to energy efficiency, and almost all of them involve tweaking weather forecasts. Rolnick thinks its development is in the hands of researchers in the short term.
Artificial intelligence will allow improving the structure of buildings to minimize lighting, heating and air conditioning costs
Artificial intelligence will make it possible to predict energy demand, which largely depends on local weather, and also improve the structure of buildings. thus minimizing lighting, heating and air conditioning costsdepending on the time of year. This, added to satellite imagery studies, will allow us to prioritize any remodeling tasks that need to be carried out on properties. For example, a project at the Borderstep Institute in Berlin showed that applying predictive machine learning algorithms could heat a block of 250 flats. at least 20 percent less energy.
Others focus on industries that top the list of top polluters. About transportation (artificial intelligence will make it possible to group shipments, minimizing the number of trips and therefore optimizing logistics routes) and agriculture (Algorithms will show which crop will be best in a given area, what is the best way to regenerate the soil, how fertilizer consumption can be reduced and pest control can be automated). In the case of industry, this includes the development of new materials and the creation of more efficient clean energies.
Reduce emissions by 4% in 2030
Apart from the undoubted benefits to ecosystems, this It will have a positive impact on the global economy. According to data obtained from the report prepared by PwC, the consulting firm commissioned by Microsoft, the use of these latest technologies in environmental cases could contribute $ 5.2 billion to global accounts by the end of this decade. Reduce greenhouse gas emissions by 4 percent by 2030. The most optimistic estimates, such as BCG Gamma, put this rate in the 5 to 10 percent range. So artificial intelligence has the chance to be the long-awaited middle ground between economic growth and sustainability.
Extreme events have increased in recent years as a direct result of climate change. Hurricanes, heavy rains and devastating fires They are becoming increasingly common and can occur at almost any time of the year. Artificial intelligence will also hold the key to better managing this problem. Solutions are diverse and can be implemented before, during and after these disruptive events. “For example, autonomous drones guided by artificial intelligence can be used to prevent fires or search for survivors in floods and earthquakes,” says Nuria Oliver. These will also improve monitoring of deforestation.
Manage forest areas in cities
In this regard, a few months ago Google Tree Canopy project, which aims to control and improve the tree cover of major cities. «Trees are increasingly seen as a solution to reduce street temperatures and improve the quality of life of residents. But many cities don’t have the resources to know where they are or where they are located. areas where new planting efforts are most needed» explains Adam Elman, the technology giant’s sustainability officer for Europe, the Middle East and Africa. In Spain, Vitoria, Barcelona and Zaragoza already use this technology.
But not everything is rosy. Nuria Oliver explains how these technologies work still requires large amounts of energy. And as a result, it can cause high levels of greenhouse gas emissions, depending on its source. The important thing is to analyze the costs and benefits of this revolution that is called to usher in a new era.
‘B’ SIDE: HIGH ENERGY CONSUMPTION
Artificial intelligence (AI) could result in a significant energy footprint once this technology reaches its full development, which is expected within a few years. This consumption can be as follows: It will exceed the energy demand of some countries such as Argentina, the Netherlands or Sweden.
Generative AI, which can generate text, images or other data, including OpenAI’s ChatGPT, has seen rapid growth since 2022. Training these artificial intelligence tools It requires feeding models with large amounts of data, a process that consumes a lot of energy. Hugging Face, a New York-based AI development company, reported that its multilingual text-generating AI tool consumed approximately 433 megawatt hours (MWH) during training, which is enough to power 40 average American homes for a year.
However, the energetic impact of artificial intelligence does not end with education. A study published by Alex de Vries, from the DigiEconomist platform and a PhD student at the Free University of Amsterdam, found that when the vehicle starts working, every time it produces text or images, it also Uses significant amounts of computing power and therefore energy. For example, running ChatGPT can cost 564 MWh of electricity per day.
Global AI-driven electricity consumption could increase by 85 to 134 TWh annually by 2027Artificial intelligence is based on the prediction of server production. This amount is equivalent to the annual electricity consumption of countries such as the Netherlands, Argentina and Sweden.
………………………………………………..
REPORT. Darío Redolat, meteorologist at the Climate Research Foundation (FIClima)
«Professional supervision still necessary»
Although their true potential has not yet been realized, meteorologists predict that these technologies will be of great importance in the future: They will improve weather forecasts and risk situations.
-What applications does artificial intelligence have against climate change?
-Can be used in real-time monitoring of water and electricity management, monitoring of agriculture or air conditioning in buildings, infrastructure or transportation. It makes it possible to improve risk prediction thanks to the real-time information that can be obtained from weather forecast systems, but this is not only it. It can also improve the carbon footprint by optimizing efficiency and savings compared to traditional systems that are more rigid and do not have high predictive capacity.
-At what stage is the investigation?
-[La IA] It has been practiced for over 20 years but is at a more basic level. Neural networks and deep learning can be found in various fields, but not yet on a large scale. The ability of these tools to explain complex phenomena that we cannot relate or are unaware of predicts their increased application in the future.
-It can also be very useful in predicting natural disasters…
-The ability that artificial intelligence technology offers to process large amounts of complex data is truly immense. This makes it ideal for the development of monitoring and warning systems. In this way, coupled with increasingly accurate meteorological models and tracking and monitoring tools with higher spatial and temporal resolution, these will make us increasingly better and more informed. It will be difficult to achieve absolute certainty, but predictions are improving every day.
“The ability of AI technology to process large amounts of complex data is truly overwhelming.”
-Will artificial intelligence change the way we predict the weather?
-I am confident in it and hope that its development will open up many opportunities and allow the support of many areas of research. Perhaps visualization of impacts will be quicker and clearer by the way we can monitor situations that may pose risks in the short or near term, rather than by the way we predict the weather. Weather models and meteorology itself involve chaos theory; hence, even considering the leverage that AI algorithms and methodology can bring to improving models, they will always have a margin of error.
-Will artificial intelligence replace meteorologists?
-No, at least in the short and medium term. It still needs years of development and the incorporation of deeper learning models. It is currently considered a support tool where supervision by a professional is required to interpret and verify data. Automation in weather forecasting is probably easier to grasp in the medium term, but not in pure research. They lack the imagination necessary to create new ideas or paradigms on their own.
……..
Contact address of the environmental department:[email protected]