New AI model aims to forecast weather and beyond in seconds
Aurora, Microsoft’s new AI model, is speeding up 10-day weather forecasts — and it might soon predict everything from pollution to energy demand.
Rebecca Dzombak reports for The New York Times.
In short:
- Aurora is a lightning-fast AI model trained on traditional physics-based forecasting data, capable of producing accurate 10-day forecasts faster and at finer resolutions than existing systems.
- Unlike typical weather models, Aurora is designed to be multi-purpose, able to forecast other environmental factors like air quality or ocean wave height with minimal reprogramming.
- Despite the speed and potential, experts stress that human oversight and physics-based models are still essential to ensure reliability and prevent AI from generating flawed predictions.
Key quote:
“I’m most excited to see the adoption of this model as a blueprint that can add more Earth systems to the prediction pipeline.”
— Paris Perdikaris, University of Pennsylvania professor and former Microsoft researcher who led development of Aurora
Why this matters:
Imagine a model that can anticipate a smog surge days in advance or tell utilities when and where a heat wave might trigger blackouts. Better, faster forecasts mean more lives saved in extreme weather and smarter planning around health-affecting factors like air pollution. Forecasting however, isn’t just about feeding data into a black box and waiting for answers. Scientists are clear: Physics still matters, and human oversight is non-negotiable. And with cuts to federal agencies like the National Oceanic and Atmospheric Administration, the National Science Foundation, and the National Weather Service, the very data sets AI forecasting depends on could be in jeopardy.
Read more: Ripe for disaster declarations — heat, wildfire smoke and death data
