Bright
← Living questionsRECORD / Earth

From research to the forecast desk.

ECMWF brought its Artificial Intelligence Forecasting System into operational service.

Original sources ↓ · Revision history ↓

Deployed · source published 2025-02-25

The human problem

Forecasters need reliable tools beyond research demonstrations.

The prior constraint

ECMWF used its established physics-based forecasting system.

AI’s actual role

AIFS added machine-learned predictions to the operational toolkit.

The documented result

AIFS began running alongside the physics-based IFS on 25 February 2025.

Why it may matter

Forecast users gained an additional operational source of guidance.

Limitations

Deployment is not a guarantee of superiority for every variable, place, or event.

Unresolved questions

Compare performance continuously and improve forecast use.

Source history & evidence assessment
Maturity
Deployed
Claim confidence
unassessed
Event date
Not recorded
Source published
2025-02-25
Captured
Not recorded
Last source review
2026-09-05
Editorial method
Original source check
Place / relevance
Reading, United Kingdom · institution-location

Legacy source check; no named human reviewer is recorded in this projection.

Maturity describes the tested or operational setting. Confidence describes support for the particular claim; one does not determine the other.

Original sources

ECMWF’s AI forecasts become operational · institution

Institutions: ECMWF

Explore the underlying question

Related developments

Editorial connections between distinct settings and results; these links do not imply replication.

Forecasting more than one future.

A faster view of the weather.

Revision & correction history

No corrections recorded.