Artificial intelligence has become an increasingly valuable tool in addressing many of the world’s most pressing issues and could provide a useful approach in confronting the challenges of climate change by providing insights into complex environmental systems.
Several experts have noted that AI can be used to develop more accurate climate models, which can help us understand how climate is changing and predict future scenarios. This can inform policy decisions and help us develop more effective strategies to mitigate climate change and adapt to it. AI can also be employed to predict and respond to extreme weather events such as hurricanes, floods and wildfires. This can help us prepare for these events and reduce their impact on people and the environment.
AI can also be used with energy systems such as smart grids to reduce energy waste and increase efficiency. AI algorithms can help predict energy demand and supply and adjust production and distribution accordingly. This can help us reduce greenhouse gas emissions and mitigate the effects of climate change as we transition to a more sustainable energy system. Further, AI can assist in the development of new technologies for carbon capture and storage. AI algorithms, for example, can help optimise the design of carbon capture systems and predict the behaviour of underground storage reservoirs.
One example of a project designed to use AI to help solve environmental challenges is FarmBeats, which is funded by Microsoft’s “AI for Earth” programme. The initiative assists farmers to optimise their use of resources and reduce their environmental impact. Sensors and drones collect data on soil moisture, temperature and other environmental factors, and then machine-learning algorithms analyse the data and make recommendations to farmers on how to optimise their use of water and other resources.
AI has the potential to help mitigate the effects of climate change in concrete ways, but there are limitations to its application, at least at the moment. First of all, AI models require significant computational resources, which can result in high energy consumption and greenhouse gas emissions, the very thing they are being used to reduce. Developing more energy-efficient AI models and infrastructure should, therefore, be a priority.
High-quality data are, no doubt, used to make accurate predictions and recommendations. However, climate data can be sparse, incomplete, or of poor quality, limiting AI’s effectiveness. In computer lingo, if you put garbage in, you get garbage out. Collecting accurate, reliable and representative data not only leads to better AI models, but it builds trust with the public. Trust is also an issue with AI algorithms, which can be opaque and difficult to interpret. If people do not understand how AI is programmed, they are less confident in policy decisions based on the technology. To mitigate this issue, it is important to develop transparent AI systems that can be audited and explained.
One initiative that faced problems when attempting to use AI to address climate change is the “Carbon Tracker Initiative” launched by the European Space Agency to track greenhouse gas emissions from power plants. The Carbon Tracker had trouble accurately identifying and measuring emissions from individual power plants using satellite data. Countries and companies were also reluctant to share information on their emissions levels. This made building accurate models from machine learning algorithms around the globe difficult.
A 2020 report published by the Capgemini Research Institute highlights several other challenges related to the use of AI in climate action strategies. The multinational information technology services and consulting company headquartered in Paris surveyed 800 industry executives and 300 AI and climate change experts and found there is a shortage of skills and talent in the field, so organisations will very probably struggle to find the necessary expertise to develop and deploy AI solutions for climate action.
The report also points out that AI development and deployment requires significant investment in data, technology and talent. Organisations need to invest heavily in developing the necessary infrastructure and capabilities to leverage AI for climate action. Lastly, the report calls for the development of regulations and standards to ensure that AI is developed and deployed in a responsible and sustainable manner. This includes standards for data quality, transparency and ethical considerations that are not yet in place.
In April of 2023, the EU updated its proposed regulations for AI technologies to include provisions for generative AI, such as OpenAI's ChatGPT, and identifying copyright protection as a core piece of regulation. A dozen members of the European Parliament hammered out the draft legislation in just 11 days, including proposals that require companies with generative AI systems to disclose any copyrighted material used to train their models. The proposed laws may force some transparency on a secretive industry. The committee will vote on the amended law on May 11, and if successful, will advance to the next stage of negotiation.
In some circles, there is an inherent fear of AI technology, that some suspect, might arise from the fact that AI, particularly generative AI, can create content on its own, without human intervention. This means that AI could potentially create fake news or propaganda that is indistinguishable from real content, which could then be used to manipulate people or even incite violence. In addition, if AI is used in weapons systems, there is a risk that it could malfunction or be hacked, causing unintended harm or even starting a war.
There is absolutely no doubt that AI has the potential to play a significant role in addressing climate change by improving our understanding of the problem and developing more effective solutions. But it is important to ensure that AI development is done in a sustainable way to avoid exacerbating the problem.