To the casual observer, the Sun appears to be a constant, never-changing presence.
In reality, however, the Sun is a turbulent mass of plasma—electrically charged gas—continuously influenced by its magnetic field. The unpredictability of solar activity remains one of the key challenges for modern solar physicists.
One particular challenge is predicting the impact of coronal mass ejections, which come with varying levels of uncertainty. But perhaps machine learning algorithms could have provided us with more warning!
A new paper suggests that algorithms trained on decades of solar activity detected all the signs of heightened activity from the region known as AR13664 and may prove useful for forecasting future eruptions.
Coronal Mass Ejections, or CMEs, are immense bursts of plasma ejected from the Sun’s corona into space due to disturbances in its magnetic field. These violent events are often linked to solar flares and occur when magnetic field lines abruptly realign, unleashing enormous amounts of energy.

CMEs can travel at speeds of a few hundred to several thousand kilometers per second, sometimes reaching Earth within days if they are directed our way.
Upon arrival, they can interact with our planet’s magnetosphere, triggering geomagnetic storms that may disrupt satellite communications, GPS networks, and power grids. They can also intensify auroral activity, producing spectacular displays of the northern and southern lights.
Forecasting such events with precision and understanding their impact on Earth’s magnetosphere has long been a challenge for astronomers.
In a study led by Sabrina Guastavino from the University of Genoa, a team of astronomers tackled this challenge using artificial intelligence. They employed AI technology to analyze events related to the May 2024 solar storm, including flares from the region designated 13644 and associated CMEs.
The storm unleashed extreme solar activity, including a powerful X8.7-class flare!

By leveraging AI, the researchers directed machine learning algorithms to analyze vast amounts of previously collected data, uncovering intricate patterns that conventional methods might have missed.
The 2024 event provided a rare and valuable opportunity to assess AI’s ability to predict solar activity. The primary objective was to forecast solar flare occurrences, their evolution over time, CME production, and ultimately, geomagnetic storms on Earth.
When tested against the May 2024 event, the results were remarkable. According to their paper, the prediction demonstrated “unprecedented accuracy in the forecast with significant reduction in uncertainties with respect to traditional methods.”
Additionally, the forecast of CME travel times to Earth and the onset of geomagnetic storms was strikingly precise.
The implications of this study are significant.
Power grid failures, communication disruptions, and satellite malfunctions pose serious challenges when CMEs reach Earth. The integration of machine learning AI into solar activity prediction marks an exciting advancement. And for those of us who enjoy watching the sky, we may also get much-improved aurora forecasts!