Signal post, September 2021.
According to the recently published IPCC Climate Change report “from a physical science perspective, limiting human-induced global warming to a specific level requires limiting cumulative CO2 emissions”. The report highlights that “it is unequivocal that human influence has warmed the atmosphere, ocean and land and that widespread and rapid changes in the atmosphere, ocean, cryosphere and biosphere have occurred”. These damages are for instance dryness and floods, extreme weather conditions, and after all difficulties to produce food to people and increasing waves of refugees.
What has AI to do with Climate Change? This topic is thoroughly discussed e.g. in AI for good AI and Climate Science presentation series. Similar thoughts are presented also in an article called Tackling Climate Change with Machine Learning where various scientists list machine learning ways to tackle climate change. At first AI can be used to model climate change. There is still lots of environmental information that is unused. With AI technology this unused information may be applied, and the knowledge of climate change may increase. The World Resources Institute Data Lab is using AI to build precision tools and solutions that will enable critical advancements at a larger scale regarding tracking the drivers of climate change, analyzing risks, which can shift emergency preparedness and taking action to a new level in the future.
A huge potential exists in using AI for improved effectiveness to support timely and adequate decision-making for reducing CO2 emissions and mitigating climate change. Such systems exist for instance with regards electricity systems, transportation, buildings and cities, industry, and farming. AI can also intensify carbon markets or logistics. Machine learning can also enable automatic monitoring through remote sensing. Deforestation or forest fires can be monitored with this information.
The creative way of processing information with AI technology may also help us in identifying new materials which then can be innovative in preventing CO2 emissions for example in packaging or in new battery technology for improving renewable energy applications. AI may also help in weather forecasts, which then can make renewable energy production more predictive.
The ultimate solution for mitigating climate change by reducing CO2 emissions, and also the adaptation into climate change, needs individual actions and collective decisions. AI can be a tool to these actions too by visualizing the climate change processes and trends and, in that way, making climate change more understandable to citizens, and increase the motivation for individual climate actions. AI may also allow people to participate into collective discussions and also involve them more actively into affecting the decision-making processes via social media applications.
What about ethics and responsibility in applying AI within climate change actions – are these relevant? Global Partnership on Artificial Intelligence (GPAI) Responsible AI Working Group has launched a project “A Responsible AI Strategy for the Environment” whose final report will be published in the end of September 2021. Based on this report an action-oriented roadmap on the intersections between responsible AI and climate action will be disseminated in November 2021. The roadmap will include a mapping of the key areas that would benefit from government investment in public research, an analysis of what the responsible use of AI in the context of climate action entails, including the risks in terms of privacy, fairness, control, safety, and security, a proposed capacity-building roadmap, and a public research roadmap identifying key areas requiring government support because of lack of short-term incentives.
Apparently, almost all the ethical issues covered in AI ethics also entail with AI in climate change use. Creating AI technology, in other words machine learning algorithms to process data, ethical issues should be considered, although the primary goal to mitigate and manage climate change is inherently recognized as responsible action. The possibility to misuse data, ignore privacy issues and manipulate people still exist.
What is more, from a future perspective, the way climate change issues are handled today will affect the next generations, and this turns it into a higher-level responsibility action. The attitude towards climate change knowledge could be hidden in the machine learning algorithms, if not expressed transparently. Also, the end users, which are e.g. citizens, should have enough knowledge to understand both the basics of climate change as well as machine learning to have potential to reach the knowledge produced in AI processes. This places a different kind of responsibility to us all as active inhabitants of our planet Earth: each of us need to take responsibility to increase literacy and gain knowledge about AI and climate change, express opinion, take action, get the voice heard and find opportunities to not only observe but participate in decision-making for our good future and for ensuring a good future of those coming after us. Every citizen, however, does not have capabilities to follow and understand what is happening in climate change and in AI issue. Therefore, ethical use of AI, also in climate change processes, needs societal attention and lobbying.
The writers: Nadezhda Gotcheva and Nina Wessberg (VTT)