Ver-AI Workshop on Verification of AI-Based Meteorological Forecasts

Global medium-range meteorological forecasting models based on methods from machine learning, with training data provided by various reanalysis experiments, have seen rapid development over the last few years. In terms of simple and overall measures of skill, such as mean square errors, AI-based forecasts clearly show very promising skill, but a proper and in-depth assessment of these models and the quality of their forecasts is still very incomplete. Following on from the successful VerAI workshops in 2024 and 2025, we want to continue with the discussion about how to carry out such assessments. For instance, while there is a well-established methodology for verification of meteorological forecasts, it is not clear to what extent this methodology needs to be adapted or modified for AI-based models. The workshop will allow plenty of opportunity for discussion, in particular (but not limited to) focus points, to be announced soon.