Google's New AI Model Claims Superiority Over Advanced Weather Forecasting Systems

Jakarta - Google’s DeepMind team has recently unveiled its latest AI model for weather prediction named GenCast. According to a paper published in Nature, researchers claim that GenCast surpasses the top operational weather forecasting system globally, known as ENS (European Centre for Medium-Range Weather Forecasts).

As reported by Techcrunch, GenCast offers a different approach compared to traditional deterministic weather models that tend to provide a single best estimate for future weather conditions. Instead, GenCast employs a probabilistic approach, generating a set of 50 or more predictions, each representing different possible weather scenarios.

To assess accuracy, the DeepMind team trained GenCast using weather data up until 2018 and evaluated its performance with data from 2019. The results indicate that GenCast achieved an impressive accuracy rate of 97.2%, outperforming the ENS model.

Google has stated that GenCast is part of a series of AI-based weather models currently under development, with plans to integrate the model into various services such as Google Search and Maps.

Moreover, with the introduction of this new AI model, Google aims to provide public access to up-to-date weather forecasts and historical data generated by GenCast, facilitating independent research and the development of new weather models.

AI, weather, forecasting