Given the potential for significant impacts to power systems, a better understanding of tornado frequency, intensity, and geographic distribution in a changing climate is critical. However, climate models don’t have projections of tornadoes, thus alternative methods are needed to understand how tornadoes may change in the future. This study explores the application of a novel machine learning method to reproduce historical tornado patterns and project future tornado activity. This method was found to effectively replicate past tornado trends with limited inputs. Projections from this method suggest an increase in tornado activity in the Midwest, with a shift in tornado timing toward early spring and winter. While this method is constrained by the ability of climate models to simulate the frequency and spatial distribution of large-scale circulation patterns, it highlights a promising method for filling critical climate data gaps.