A recent survey of 6,000 UK adults indicates that while AI is increasingly used in weather forecasting, public confidence remains significantly higher in traditional Numerical Weather Prediction (NWP) methods compared to Machine Learning Weather Prediction (MLWP).
- 87.7% of UK respondents trust traditional NWP weather forecasts
- Only 49.4% express confidence in AI-driven weather predictions
- Met Office aims to combine AI with physics-based models while addressing public skepticism
What happened
A survey involving 6,000 adults in the UK revealed that the majority still prefer and trust traditional methods of weather forecasting, specifically Numerical Weather Prediction (NWP), over emerging AI-based approaches such as Machine Learning Weather Prediction (MLWP). This reflects a general apprehension toward relying on AI models for weather forecasts despite the increasing role of machine learning within meteorology.
The study was conducted by the Met Office and published as part of their research on Artificial Intelligence for the Earth Systems. The survey measured public perception of the accuracy and reliability of AI-assisted forecasts compared to established physics-based models.
Why it matters
The confidence gap identified in the survey presents a significant barrier to the wider acceptance and use of AI in weather forecasting. Since public trust influences how people react to weather warnings and forecasts, skepticism about AI's accuracy could hinder the effectiveness of future forecasting innovations.
Met Office scientists emphasize the importance of clearly demonstrating the success and reliability of AI models alongside conventional methods to gain public confidence. Without this trust, the potential benefits of AI-enhanced weather prediction — including improved detail and longer-term accuracy — may not be fully realized or acted upon.
What to watch next
The Met Office plans further studies to track how public attitudes evolve as AI technology in weather forecasting matures and becomes more integrated with physics-based approaches. Efforts will focus on transparency and validation to prove AI’s value in real-world scenarios that matter to users.
Additionally, the blending of AI and traditional meteorological models is underway, with careful evaluation to ensure any new hybrid forecasting methods meet scientific standards before public deployment. Observers should watch for updates on how these advancements impact both forecast accuracy and public reception over time.