AI models outpace traditional tools in predicting hurricane paths, experts say

In one outlet Atlantic season, the Google DeepMind AI model accurately anticipated Hurricane Isaias's landfall near Destin, Florida, whereas traditional models initially projected a path toward southeastern Louisiana. Experts like Michael Lowry described the AI as the "gold standard" for hurricane prediction, and University of Miami researcher Brian McNoldy noted its rapid ascent to top performance after only a short period.

The National Hurricane Center incorporated the DeepMind model into its official forecasts last year and now uses a consensus approach that mixes AI and conventional models, with director Michael Brennan emphasizing the essential role of human interpretation. While the AI excels at track forecasting, it still underperforms in intensity prediction, prompting ongoing feedback to improve the system. The technology's fast update cycle promises quicker advances than the annual updates of traditional models, though forecasters remain uneasy about its black-box nature.