Join the Geospatial Forum for "Geospatial Machine Learning in the Applied Geosciences: Pitfalls and Promise in 2026." Corey Scheip, Ph.D., a senior geoscientist and lead for BGC Engineering's Data Science and Earth Observation team, will discuss the state of geospatial machine learning in the applied geosciences, focusing on landslide susceptibility and inventory mapping as two examples. Susceptibility modeling is unusually easy to do badly: DEMs, soil surveys, and landslide databases are easy to download, and a respectable-looking AUC is a few lines of AI-generated code away. But are the results useful? Models for mapping landslides with lidar, on the other hand, are a different beast. They are more complex to create, but done with care, they can 100x the speed of human mappers and augment the way we’ve been mapping landslides for decades. In 2026, with modern AI-driven tooling, the promise for both applications is real. However, most of the promise doesn't come from the models themselves – it comes from the underlying data and decisions that the human expert made along the way. 

You can attend Scheips's talk in person at 5103 Jordan Hall or watch it live.