Aarva

The Conversation ·For the curious

Why AI Has Trouble Predicting the Fury of Hurricane Intensity

by Chanh Kieu, Associate Professor of Atmospheric Science, Indiana University

Published 2026-09-23 12:44:25+00:00

If a hurricane is naturally chaotic, can a machine ever predict how strong it will get?

0:00 / 8:07 · Narrator Sulafat

Context

Global artificial intelligence models can now predict the weather with startling accuracy, as long as the view remains wide. But when a storm approaches a coastline, that certainty often breaks down. An essay published on 23 September 2026 pauses on this exact blind spot: the difficulty of predicting how fast a hurricane will intensify. The problem extends far beyond a lack of detailed radar data over the open ocean. The real friction comes from the atmosphere's natural chaos, forcing a hard look at whether machines trained to find perfect patterns can ever model a storm defined by unpredictable fluctuations.

Show notes

An analysis of why artificial intelligence struggles to forecast rapid changes in hurricane intensity. The problem stems from two main obstacles: a lack of complete, three-dimensional weather data over open oceans, and the inherent chaos of storm development. Improving these systems will require forecasters to abandon single-number predictions in favor of probability models that account for unpredictable fluctuations.

Read on The Conversation →