We can agree that wildfire is natural in western forests, and at the same time not want fires to burn down our houses. Some hard lessons have been learned about fire behavior in recent years.
It would certainly help if fires were more predictable, and they just might be. Researchers at the University of California-Irvine used machine learning, a kind of artificial intelligence, to predict how big a wildfire might get, based on the ignition point and the conditions where the fire started. The tests have shown some promise in predicting the eventual size of fires; research co-author Shane Coffield visits with details of the work.