Landslides can be caused by a variety of factors, from massive rainfall to ground motion and slope. And though AI has already been deployed in landslide prediction models, geologists don’t always know which of those variables were most influential in the machine-generated predictions.

To make landslide prediction more accurate, the researchers relied on a type of AI that makes the interactions of a variety of complex factors clearer. They trained the AI on data from the Himalayas, which are susceptible to landslides because of earthquakes, steep slopes and extreme weather.

The new AI tool was about as accurate as other AI models now in use, but the researchers say their model takes less computing power than others and allows its human operators to learn more from its analyses.

“Similar to how autopsies are required to determine the cause of death, identifying the exact trigger for a landslide will always require field measurements and historical records of soil, hydrologic and climate conditions, such as rainfall amount and intensity, which can be hard to obtain in remote places like the Himalayas,” Seulgi Moon, an associate professor of Earth, planetary and space sciences at UCLA and one of the study’s co-authors, said in a news release.

The researchers say their new model might also help researchers predict other events, such as wildfires and earthquakes. They add that the technology could improve prediction in places such as California, where recent landslides probably fueled by rainfall have caused luxury homes to slide into a Los Angeles County canyon.

The jury is out on whether, or when, AI could become the primary tool used by meteorologists or geologists, but the technology is progressing in a variety of fields. For those affected by landslides and other natural disasters, that future cannot come fast enough.