Possessing immense predictive power, artificial intelligence (AI) and machine learning (ML) have turned into buzzwords over the course of the last decade. At present, AI is being used to revolutionize renewable energy supply and make energy markets more efficient. However, less attention has been focused on the climate security applications of AI. Even with an AI-driven transformation of our energy systems, it is unlikely the world will be able to keep global temperatures from rising 1.5 to 2 degrees C.
The global community must lean forward and adopt strong adaptation measures to prepare for the effects of climate change, which will worsen with time. Fortunately, AI is perfectly suited to augment climate preparedness measures. AI offers four capabilities that can aid the adaptation process.
AI provides more reliable prediction methods.
AI’s ability to harness big data and quickly identify patterns in this data makes it an ideal prediction tool. There are several opportunities to leverage AI’s predictive power to bolster climate security.
First, AI can be used to predict the likelihood and severity of extreme weather events, such as hurricanes, wildfires, and drought. AI already assists researchers in achieving 89 to 99 percent accuracy identifying tropical cyclones and weather fronts.
Second, AI can also be used to tune current climate models to better-forecast sea level rise and atmospheric heating. Climate assessments typically rely on downscaled global-scale climate models for information at the regional and local levels. With this downscaling often being an imperfect process, AI offers the ability to pair these downscaled models with more local data (insurance claims, weather reports, migrant levels, etc.) in order to build a more accurate model.
Finally, AI’s predictive power can assist in saving wildlife and biodiversity. The Protection Assistant for Wildlife Security (PAWS) project is a shining example, with the organization employing a machine learning model to predict where poaching may occur in the future in order to protect endangered species.
AI enables better monitoring of climate trends.
AI helps researches and government officials keep tabs on the determinants and effects of global warming. In China, IBM’s Green Horizon project uses an AI system to track pollution sources and recommends courses of action to best mitigate smog levels (close power plants, limit drivers, etc.). Computer vision capabilities offered by AI can also monitor deforestation efforts and the state of wildlife. Long Live the Kings, an organization based in Washington State, is also leveraging AI to track salmon movements and build models that will help support habitat protection.
AI enables more robust planning and coordination.
With limited resources, policymakers will have to prioritize which regions and economic sectors warrant government climate protection resources. Capable of parsing through a large portfolio of investment options, AI can aid policymakers in these tough decisions while helping the nation avoid the misallocation of funding that can stem from large federal grants.
AI can help the global community create a more sustainable and resilient economy.
AI offers tools that can make critical sectors of our economy more resilient to the fallout of a changing climate. AI has enabled “Smart Farming,” innovative farming techniques that leverage data and emerging technology to farm more efficiently and produce more resilient crops. In India, AI has allowed farmers to achieve 30 percent higher groundnut yields per hectare of land by equipping them with information on how to apply fertilize and the optimum sowing dates. At the same time, the predictive capabilities of AI allow farmers to better predict when drought may occur, allowing them to make the necessary preparations ahead of time.
AI can also harden the power grid. The Department of Energy’s SLAC National Accelerator Laboratory, for example, is utilizing machine learning to identify vulnerabilities in California’s power grid and strengthen them before failures occur.
AI is not without risks.
There is no doubt that AI offers a promising way ahead as we seek to adapt to climate change. However, like any information technology system, AI can be hacked. Complex systems put in place to provide early warning of extreme weather or optimize power grid output are susceptible to maligned actors. Were one of these systems to go down, the consequences could be devastating.
AI is also difficult to interpret. Far more complex than other predictive statistical methods, AI and ML techniques often feel like black boxes, making it difficult for AI practitioners to describe the intricate decisions made by the algorithms to a curious public. Were an AI system to fail and result in the loss of life, such an opaque decision-making process could be difficult to defend.
AI-based decisions are not free from the biases that come with any human-made decision. Every model requires making certain assumptions, and there is a real risk that some AI models may be biased towards certain notions, socioeconomic groups, or ethnic makeups.
Finally, the power of AI can often breed overreliance on the technology. Although AI offers an immense amount of predictive power, AI-based models are never perfect. There is a distinct possibility that government officials become too trusting of AI and fail to verify the veracity of the predictions it makes. In the realm of climate change, this complacency could lead to officials missing extreme weather events, wasted resources, and even loss of life.
Despite these risks, AI is a powerful and versatile tool with real applications for climate security, both in terms of preparedness and response. From managing refugee flow after an extreme weather disaster to controlling damage to infrastructure caused by rising sea levels, AI will prove a pivotal tool after climate-related disasters occur as well. Though the future of our planet can oftentimes look bleak, climate-centric AI technology offers a promising way ahead.