Climate change is one of the greatest global challenges. Even though artificial intelligence and machine learning entail a significant carbon footprint, they could be used to reduce greenhouse gas emissions.
The effects of climate change, along with the resulting changes in ecosystems and the degradation of natural resources, are becoming increasingly apparent. The 2018 report of the United Nations Intergovernmental Panel on Climate Change (IPCC) identifies the disastrous consequences of inaction or delaying action aimed at eliminating greenhouse gas emissions.
Harnessing and mitigating climate change, as well as preparation for the inevitable consequences of global warming pose multifaceted and multidimensional problems. However, artificial intelligence (AI) and machine learning could prove helpful in solving them.
New data processing and data analysis technologies could be used to reduce greenhouse gas emissions, among other things, through changes to electricity systems, transportation, buildings, industry, land use and green investment. Additionally, they could be useful in climate modeling, predicting climate-related risk, and managing climate disasters.
The possible applications of AI and machine learning could include the use of satellite images for improved monitoring of deforestation, as well as the development of new materials that could replace steel and cement, the production of which accounts for 9 per cent of the total global greenhouse gas emissions.
AI has the potential to transform entire industries. This applies in particular to electricity systems, which are responsible for about a quarter of the man-generated greenhouse gas emissions. Many electric energy systems collect huge amounts of data, which is why machine learning could help in forecasting electricity production and demand, as well as reducing energy waste in buildings.
Artificial intelligence can also promote green financing through the analysis of data and information made public by companies, as well as the use of natural language processing algorithms for the identification of climate risk.
“Green” investment opportunities relate to investing in sustainable technologies. Financial institutions invest in this way, among other things, by developing “green” financial indexes, which focus on low-carbon energy, clean technology and environmental services. They also design carbon-neutral investment portfolios that exclude companies with relatively high carbon footprints.
In certain market sectors such investment strategies lead to major shifts towards renewable energy alternatives which are seen as having greater growth potential than traditional energy sources such as coal, oil and gas. This applies, for example, to the public utilities sector.
The development of green investment strategies can be facilitated by the application of deep learning algorithms — both for portfolio selection, based on the features of the stocks involved, as well as for the timing of the investment, using historical patterns to predict future demand. This could help in maximizing both the impact and the scope of climate investment strategies.
High cost of climate change
It is estimated that the combined cumulative effect of climate change on global financial assets amounts to approximately USD2.5 trillion. However, it is difficult to predict the long-term consequences of global warming, because it is not certain where, when or how it will affect a given company’s share prices, or the level and valuation of a given country’s public debt. Additionally, it is difficult to make predictions due to the planning horizon in the financial system — most financial analysts and investors focus on the analysis, risk evaluation and forecasting of potential profits in the short term: on a quarterly or annual basis.
Therefore, such an approach does not take into account the medium- and long-term risk, which includes the physical impact of climate change on the assets or the distribution chains, the legal effects of these changes for profit generation, and the indirect market consequences affecting the demand and supply dynamics.
One possible solution could be predictive climate analytics, which aims to anticipate the financial consequences of climate change. It involves analyzing investment portfolios, funds and companies for the purpose of identifying areas and entities characterized by increased exposure to risk caused by climate change. Such entities at risk may include, for example, enterprises engaged in wood processing, which could go bankrupt as a result of numerous forest fires, or organizations involved in the extraction of water, the sources of which could become contaminated as a result of landscape changes.
In the context of risks to the financial system, predictive analytics of climate change — in addition to the possibility of using deep learning algorithms — could also involve, among others, the development of arbitrage strategies utilizing the risk related to the carbon footprint in the financial markets. It could also utilize machine learning to forecast carbon emission prices or to analyze the evolution in media coverage of climate change, in order to enable dynamic hedging against the climate change risk.
Artificial intelligence and machine learning could therefore play an important role, at least in mitigating the effects of climate change, including its financial effects. However, this requires us to become aware of the need for joint action and collaboration throughout the value chain of the investment. At the same time the individual institutions are responsible for managing climate-related risks and opportunities. The ones that are proactive, are able to create value for their customers and partners, achieve a competitive advantage, reduce systemic and financial risk, and make a valuable contribution to the society as a whole.