Preparing for the future: How AI is reducing catastrophic impacts caused from climate change

During Earth Summit 2021, we invited experts and leading innovators to discuss the developments, opportunities, and current limitations of artificial intelligence (AI) in our Fire track. The two exciting panelist discussions were ‘The Smart Grid: Harnessing the Power of AI’ and ‘Wildfires in California and Beyond: Monitoring for Risk with AI and Machine Learning’. In these discussions, we explored how AI, machine learning, and other technologies can reduce or avoid catastrophic impacts caused by climate change. It was agreed that now, more than ever, we need to advance our technology and reduce our reliance on outdated systems to ensure sustainability and preparedness for the future.

The recent winter storm in Texas and the subsequent negative consequences are key examples of why we need to develop and adopt AI to protect our communities from the impacts of climate change. On a short-term basis, AI can create energy supply forecasts – such predictions would have indicated to Texan utility companies the need to increase their backup power supplies. On a long-term basis, AI could also be beneficial in making long-term predictions; this would assist utility organizations in making informed decisions on which future infrastructure investments to make.

Investing in infrastructure, like energy storage for electrical grids, is another example where AI can enhance the way we manage our energy needs. Our grid system is a bi-directional system where consumers can both draw energy and produce energy. This, along with the fact that intermittent renewable sources such as wind and solar may not be available all year round, demonstrates that investing in energy storage infrastructure is essential to avoid reliance on fossil fuels. 

Furthermore, AI used in energy grids would allow utility organizations to forecast electricity consumer demands (outputs) and consumer supply (inputs) to better optimize the operation of the storage system. This will enable energy providers to flatten the load during peak times and be able to use low carbon or zero-carbon resources stored in batteries more effectively. ‘Grid integrated buildings’ is an example of controlling energy usage based on supply and demand of energy levels throughout the day.

“By controlling 25% of the building’s energy, you can reduce the energy use of that building by 30 or 40%”. 
- Dr. Jeremy Renshaw, Senior Program Manager at Electric Power Research Institute.

Predictive modeling through the use of satellite imagery is an important tool in risk mitigation, which can be used for both predicting energy loading on an electrical grid and detecting and predicting megafires. “Satellite imagery is one the best things we can potentially do in the future,” Dr. Jeremy Renshaw.

In regards to fire management, predictive modeling can help detect fire that ignited only 15 minutes earlier. Through advanced satellite sensors, fires can be detected through clouds and up to 22,000 feet above sea level. Furthermore, small government-owned satellites are currently being used to monitor Californian forests by mapping annual vegetation data in high resolution, including fire fuels and overall forest structure. This AI technology can even map the height of vegetation. Interestingly, this information can be used to drive wildfire modeling, for example, identifying trees across California’s utility grid that are at higher risk of falling onto power lines and igniting wildfires. Utilities spend a billion dollars every year in forest management, and advanced satellite imagery could be highly beneficial in identifying the highest priority areas. 

While there are some considerable advances in technology assisting these fire management and energy grid industries, there are still gaps in the adoption of AI. One of the reasons for this is data management. Sharing data is a challenge that affects a diverse range of sectors. Currently, there are enormous amounts of datasets that exist in, unfortunately, very fragmented and customized formats. Data creators need to make data accessible and freely available for industries to advance using AI. However, this comes with some caveats, such as the need for a robust chain of data custody, detailing where the data originated, what amendments have been made, and by whom. 

As reliance on these data sets and AI technology increases, we need to consider the security of this information. While unlikely, an attack on AI has huge consequences. For example, a foreign attack on a power grid would have the capacity to remove all functions from an energy grid, causing life or death consequences for those who rely on the supply. Along with advanced AI, it is equally important to have good political relationships as a protective mechanism. 

Regardless of the issues associated with security and data sharing in AI, there needs to be more individuals and organizations (in general) working on solving climate change.

“Honestly, we need all types of talents, all people from different backgrounds from all over the world. This should be all hands-on deck moment. This decade is critical to getting things moving on climate change, particularly in technology.”
- Dr. David Marvin, Co-founder and CEO of Salo Sciences, Inc.


To maximize the benefits of AI, policymakers and organizations need to invest in infrastructure to prepare us for the future better. In the case of AI assisted wildfire modeling, one such option would be to reduce the current level of biofuel throughout Californian forests. This could be completed by recycling a fraction of California’s urban wastewater to irrigate forests. This will keep the forests green and remove the fire fuel loading, providing an opportunity to reset and better prepare for long-term fire management and planning.

In the case of the electricity grid, the United States currently has more power outages than any other developed nation. There is an opportunity for infrastructure to be upgraded and AI to be implemented into the grid to future-proof our energy requirements and make our energy supply more resilient, particularly with the increasing need to utilize renewable sources. 


Wildfires in California and Beyond: Monitoring for Risk with AI and Machine Learning

Guest speakers: Dr. David Marvin, Co-founder and CEO of Salo Sciences, Inc. Kian Mirshahi, CEO of Dr. Natasha Stavros, Director of Earth Lab Analytics Hub Allison Wolff, CEO of Vibrant Planet

The Smart Grid: Harnessing the Power of AI

Guest speakers: Josh Lehman, Head of Product Management at Stem, Inc. Kristina Libby, Chief Science Officer at Hypergiant - @KristinaLibby Kerri Devine, Director of Engineering at Arcadia - @KerriDevine Dr. Jeremy Renshaw, Senior Program Manager at Electric Power Research Institute (EPRI)

About the Speaker

Want to get more involved?

Volunteer or  Donate!