A group of students from St Mary's Catholic School and Fellows Kendra Gilmore and Paloma Jol from Northumbria University, set out to investigate whether artificial intelligence (AI) could be used to predict geomagnetic storms. They used a machine learning technique called recurrent neural networks (RNNs) to analyse data on Earth's magnetic field and solar wind conditions.

The students trained their AI model on a dataset of historical data. They then used the model to predict future geomagnetic storms. When the model was trained on 40% of the data, the predictions were inaccurate and did not resemble the actual values. However, when the model was trained on 90% of the data, the predictions became more accurate, with some correlation between the predicted and actual values.

While the results were promising, the students concluded that AI is still not reliable enough to accurately predict geomagnetic storms as even with 90% of training, the results were not particularly accurate. They suggest that further research is needed to improve the accuracy of AI models for space weather forecasting.

This student-led research highlights the potential of AI to revolutionise our understanding of space weather. By continuing to develop and refine these models, we may one day be able to accurately predict and prepare for the impacts of geomagnetic storms!