Unlike traditional forecasting models that rely on physics equations and real-time data like temperature and wind shear, AI models use neural networks to analyse historical storm data and atmospheric conditions. These algorithms mimic the human brain to identify patterns and make predictions. However, AI models are sometimes referred to as 'black boxes' because their decision-making processes are not always transparent.