Can AI help to battle climate change?
23 April 2024
By Dr Alexander Los, IHS Senior Specialist in Urban Environment and Climate Change
Cities are growing, and so are their greenhouse gas emissions. Already in 2001, the IPCC indicated for the first time that “the balance of evidence suggests that there is a discernible human influence on global climate”, a finding for which evidence has only been increasing. Our society flourished as energy became available in abundance. This circumstance incidentally enabled scientists to understand the climate system well enough, creating evidence of the negative consequences of our undamped energy consumption.
“The longer we wait to reverse climate change, the more swiftly we have to transform our fossil fuel-based regime into a fair and climate-neutral socioeconomic system.” says Dr Alexander Los, IHS Senior Specialist in Urban Environment and Climate Change.
Exploration of new opportunities
Motivated by the urgency to mitigate and adapt to climate change, climate scientists begin to explore new opportunities of machine learning and artificial intelligence (ML/AI) to accelerate and optimise their research. For example, ML/AI allows to overcome limitations in state-of-the-art climate modelling to discover tipping-points in the climate system. These technologies are used for disaster prevention to forecast specific impacts of an impending climate extreme event. ML/AI-based large-language models, feeding popular chatbots, can respond deliberately and inclusively to sceptical attitudes towards climate change, which prompts users to revise their climate-sceptical stances.
Use of AI for climate mitigation and adaptation
In climate research the “physics-informed machine learning” technology forces ML/AI models to obey fundamental governing laws of the climate system, which is an important constraint to make such models better suited for the climate process being studied. This rapidly developing technology seeks to increase data efficiency, to accelerate the required training process for the models, to improve prediction reliability for new scenarios, and to enhance transparency and interpretability that make climate models more trustworthy. Combining ML/AI technologies with modern earth system models has therefore a high potential to provide more physically consistent and scientifically sound predictive climate models.
Merging the power of technologies
As global greenhouse gas emissions are still growing, governments and society must embrace more transitional approaches to mitigate climate change and realise that uncertainties in scientific outcomes do not reduce confidence in the necessary mitigation and adaptation actions. On the contrary, the risk for humanity of waiting to mitigate climate change is likely to exceed the risk of the mitigation actions themselves. In this risk-acceptance process, the role of ML/AI will become more eminent in the effort of quickly gaining more insights into the necessary inter-sectoral relationships and to accelerate the transformation of unsustainable practices. Can AI help to battle climate change? It looks like it can by merging the power of both technologies to mitigate and adapt to climate change more effectively, purposeful and equitable.
Brought to you by:
