AI and climate change
Last March, the Copernicus Climate Change Service (C3S), one of the six thematic information services provided by the Copernicus Earth Observation Programme of the European Union, announced that the world’s average temperature in February 2024 was the highest among February temperatures in history.
The C3S further announced that the 12-month average temperature for the period running from March 2023 was 1.56 degrees Celsius higher than the pre-industrial average, marking an unprecedented record in history.
The record-high temperature went over the 1.5 degrees Celsius increase, which was the last borderline that the parties to the U.N. Climate Change Conference held in 2015 promised to adhere to by adopting the Paris Agreement.
Korea is not an exception to global climate change. According to the “2023 Abnormal Climate Report” released by the Korea Meteorological Administration on April 29, the average annual temperature in 2023 was 13.7 degrees Celsius, the highest since records began in this country.
The data explain a lot of abberations we had gone through in 2023: the tropical night phenomenon that swamped Seoul in the midst of September for the first time in 88 years, the number of people suffering heat illnesses which was 1.8 times greater than the 2022 level, and the size of forest fires and rainfall in the southern regions during the rainy season that hit all-time highs.
Against this backdrop, AI-based technologies are rising as a next-generation solution to reduce carbon emissions, the primary source of global warming, and adapt to climate changes already unfolding on this Earth.
Just in time, the day the “2023 Abnormal Climate Report” was released, Korea held the AI-based Green Digital Transformation Conference, co-hosted by the Presidential Committee on Carbon Neutral Green Growth, the Government Committee for Digital Platform and the Ministry of Science and ICT.
The conference was designed to explore various ways to cope with climate change using AI and digital technologies and was a follow-up on the Promotion Plan of Carbon Neutrality through Digital Conversion announced by the Presidential Commission on Carbon Neutrality and Green Growth in November last year.
One of the ideas suggested at the conference was to invest more in AI technologies.
Currently, about 75 percent of AI applications are focused on customer management, marketing and sales, software engineering and R&D.
Considering that AI has a great potential to enhance productivity and bolster human capabilities, it is reasonably expected that AI would be able to assist with predicting abnormal climate change and increasing energy efficiency of industrial facilities.
Another idea suggested was to accelerate the development of novel materials to mitigate climate change by building autonomous laboratories operated by AI and robotics technologies.
Such AI-based labs may be able to replace human researchers in exploring and synthesizing novel materials. A new-structure Zeolite, for example, can be used to remove impurities and absorb carbon and to develop alternative gas to reduce emissions in manufacturing processes for semiconductors.
Some pointed out that AI-driven technologies present both pros and cons, as illustrated by data centers to support AI models. Such data centers tend to require more power and increase water demand, but they also significantly improve operational efficiency across various industries. Some commented that AI-based technologies could only supplement, rather than replace, various factors essential to mitigating climate changes, such as policy support, climate technology and climate finance.
The conference introduced various real-world projects where AI brought tangible changes to energy consumption and efficiency.
For example, AI-enabled control of individual servers within a data center saved power by up to 55 percent, virtual carbon oxide sensors along with air and fuel control for each heater optimized oxygen injection within a factory where previously the injection was made manually by observing flame colors within the heater, and an AI-based IoT platform for energy control which connected about 5.63 million devices around the world saved electricity by up to 60 percent through air conditioner compressor speed control, resulting in energy savings of 4.4 gigawatt-hours since 2022.
As the chairman of the Science and Technology Sub Committee of the Presidential Commission on Carbon Neutrality and Green Growth, my key takeaways from the discussion with various AI experts at the conference were as follows.
First, AI-based responses to reduce carbon emissions that are available under today’s technology can be summarized as facility optimization, operational efficiency, advanced monitoring and platform optimization. AI-based responses tailored more to combat ongoing climate change can be summarized as climate change prediction, social impact prediction, simulations and accident responses.
Second, AI-based autonomous labs, which can replace human researchers in certain tasks in the future, may be able to develop new materials or processes that will revolutionize combating climate change.
However, their impact is difficult to gauge at this stage since we do not have any tangible examples of such breakthroughs yet.
Third, ever-increasing AI learning and advanced specifications of AI devices are expected to drive energy demand even further, which will tighten up the supply of coal power-generated energy in the short term and even their alternatives, such as renewable energies, in the longer term.
It is still unclear whether AI will bring a clear solution to the climate changes that we have been struggling with for more than a decade.
However, I remain hopeful for AI’s unprecedented potential since we have only seen the tip of its iceberg.