Artificial intelligence and related technologies have enabled massive leaps in the fields of marketing, customer service, transportation and industrial processes. AI has even composed music. Its next challenge? Predicting natural disasters and improving the way we respond to them.
Earlier Warnings
Earthquakes are notoriously hard to predict. They strike unexpectedly and spark wildfires and tsunamis that add to the damage. Scientists have tried, but no one has yet found a way to reliably predict them. Some researchers, though, think AI might be the key to finally doing it and saving lives in the process.
A team of geophysicists from Pennsylvania State University and Los Alamos National Laboratory has been simulating quakes, gathering data and analyzing it using machine learning, which has allowed them to gather more information on more variables than ever before.
What they’ve found is that a creaking and grinding noise in the acoustical data has proven to be a useful predictor that allows them to pinpoint when an event will occur. While these findings might be difficult to translate into the real world, some scientists are optimistic that AI could finally provide us with the ability to forecast earthquakes.
Machine learning plays a role in the weather predictions we get every day, and that same technology can be used to forecast when a hurricane will hit. AI has also proven 30 percent more effective than traditional models at predicting the intensity of hurricanes.
This improved accuracy helps people better prepare for the event. A 30 percent improvement is hugely consequential, especially given that just three hurricanes, Maria, Harvey and Irma, impacted at least 26.5 million people in the U.S. in 2017.
Predicting Damage
AI is also showing promise as a way to improve disaster relief. A company called One Concern has created a predictive AI program called Seismic Concern and is also working on similar programs for wildfires, floods and hurricanes.
Seismic Concern takes data about seismic activity, the structural integrity of nearby buildings and the demographics of the affected people and analyzes it to help relief workers determine what areas need what kind of assistance.
The company has products that facilitate disaster simulations and help leaders make decisions during and after a real event, enabling users to both prepare for and react to an event.
Social and Smarter Relief
While social media might not always be accessible after a natural disaster, when it is, combining it with advanced data analytics could help improve disaster relief. After a disaster, the best information comes from the ground from relief workers and the victims themselves. AI programs can gather all the social media content related to an event into one place, analyze it and provide leaders with a better idea of the scene in the disaster area.
When combined with imagery from drones and satellites, this dataset becomes even more useful. Governments and relief groups can then use this information to identify which areas are the most badly damaged and where the people that need relief are.
Victims could also use this information to gain a better understanding of their situation and get advice on what to do. ‹ Users could then conduct a search with an AI voice assistant like Amazon‘s Alexa or Apple‘s Siri. They could pose questions about the impacts of the natural disaster, the progress of relief efforts and what actions they should take.
When relief workers, government and the public are all better informed, relief efforts may be able to take place much more smoothly and effectively.
