Researchers used AI to analyze 1,500 climate protection measures to identify those that have reduced CO2 the most so far. They focused on four sectors: buildings, electricity, industry and transport.
A team of researchers used machine learning to examine around 1,500 political climate protection measures. The aim was to identify those that have so far been most effective in reducing CO2 emissions. The study was published in the science magazine Science published.
AI identifies the most effective climate protection measures
The result: The researchers were able to identify 63 measures in 35 countries that made a significant contribution to reducing CO2 emissions – an average of around 19 percent. The most extensive reductions were associated with a combination of at least two or more actions.
In total, the 63 measures reduced 0.6 to 1.8 gigatons of CO2 equivalent. Annika Stechemesser, co-author and researcher at the Potsdam Institute for Climate Impact Research, said: “The right mix of measures is more important than a multitude of measures.”
The gradual closure of coal-fired power stations in the UK worked to some extent because it was combined with pricing mechanisms such as a minimum price for carbon. The ban on cars with combustion engines in Norway was effective because there were price incentives for electric cars at the same time.
Combination of machine learning with a statistical analysis approach
The researchers base their analysis on a database of around 1,500 climate protection measures implemented by 41 countries between 1998 and 2022. They evaluated 48 different categories.
“Previous assessments tended to focus on a narrow selection of prominent policy measures in selected countries, overlooking the hundreds of other measures,” said Stechemesser.
For their study, the researchers combined machine learning AI with a statistical approach. This enabled them to identify emission reductions in the four sectors that emit the most CO2: buildings, electricity, industry and transport.
They compared the results with climate protection measures from the database to assess which individual or combined approaches were most effective at reducing emissions. The analysis deviates from conventional approaches of reviewing the measures individually and selecting the most important ones. However, the AI made it possible to identify larger changes in climate protection measures.
According to AI: Certain climate protection measures work better in certain countries
The results show that certain measures work better in certain areas and regions – especially in combination. Price measures such as an energy tax were particularly effective in countries with a high income standard in reducing emissions associated with electricity generation.
The only single measure that would have been more effective than a mix of measures in all four sectors was taxation. According to the researchers, AI enabled them for the first time to evaluate climate protection measures in different sectors across countries. For ecologist Xu Chi from Nanjing University, the study results are alarming:
This study is a warning to countries around the world that their climate policies have had very limited impact so far. Existing policies need to be reassessed and changed.
Also interesting:
Source: https://www.basicthinking.de/blog/2024/08/23/kuenstliche-intelligenz-analysiert-kimaschutzmassnahmen/
