New research from the UK Met Office has identified a confidence gap between existing weather forecasting methods and emerging AI techniques.
The study, thought to be the first of its kind and published in the journal Artificial Intelligence for the Earth Systems, surveyed more than 6,000 UK adults on their confidence in using different types of forecasts for everyday and high-stakes decisions, as well as their perceptions of the accuracy of existing physics-based methods versus potential AI-generated forecasts.
For physics-based models, around 80% of respondents said they would be confident making weather-related decisions for daily and important events, compared with around 40% for potential AI forecasts. The confidence gap was reflected in perceptions of accuracy: around 78% rate physics-based models as accurate for daily and severe weather forecasts, compared with around 45% for AI-generated forecasts.
Dr Edward Pope, Met Office science fellow and lead author of the paper, said, “This paper deepens our understanding of current public perceptions as we approach a crucial juncture where AI-based weather models are demonstrating their potential to work alongside physics-based methods. This was a unique opportunity to compare public perceptions of established and emerging approaches to forecasting the weather. The gap in confidence and perceived accuracy highlighted in the paper demonstrates the need to clearly and transparently demonstrate the value of new approaches in ways that matter to people.”
The Met Office noted that trust has been highlighted during recent heat waves. Following a red extreme heat warning in June, post-event surveys found that 96% of people in the affected area were aware of the warning, and 97% of those took action as a result.
Although the Met Office does not currently use solely AI-based models in its operational forecasts, it has long used machine learning techniques in its forecasting process and is exploring how AI could be blended with physics-based modeling.
Prof. Kirstine Dale, the Met Office’s chief AI officer, said, “We are exploring ways of blending physics-based and AI-based modeling to deliver the forecasts that we all rely on. As with all science developments, we will robustly evaluate and validate any changes to our approach to weather forecasting before we introduce them into the model. Maintaining public confidence is critical, and this study provides us with valuable insight into what’s needed to ensure that we fully realize the benefits of the integrated approach.”
The research was supported by the UK’s Public Weather Service and overseen by the Department for Business, Innovation, Science and Trade. The Met Office said similar surveys are planned in the coming years to track changing public perceptions.
In related news, machine learning opens new possibilities for global atmospheric reanalysis