The Deep Learning Geomagnetic Perturbation, or Dagger, can accurately predict geomagnetic disturbances worldwide in advance, providing crucial time to minimise their impact.
Solar storms are a threat to Earth because they can damage satellites, cause power cuts and disrupt other technologies.
They happen when solar wind — energetic bursts from solar eruptions — hits Earth's magnetic field, which is what creates 'northern lights' and 'southern lights' in some parts of the world.
“With this AI, it is now possible to make rapid and accurate global predictions and inform decisions in the event of a solar storm, thereby minimising — or even preventing — devastation to modern society,” said Vishal Upendran of the Inter-University Centre for Astronomy and Astrophysics in India.
The model produces predictions in less than a second that are updated every minute.
Researchers have harnessed AI to look for connections between solar wind and geomagnetic disturbances.
They have been using previous instances of when a solar storm caused disturbances on Earth to help the computer pattern recognise patterns.
In 1989, a destructive solar storm caused electrical blackouts across Quebec, Canada, for 12 hours, affecting business, schools and daily lives.
The most intense solar storm ever recorded was the Carrington Event of 1859, as it sparked fires at telegraph stations and prevented messages from being sent.
Nasa said if that same happened today, the impact would be much more severe, with widespread electrical disruptions, persistent blackouts and breaks in global communications.
“Such technological chaos could cripple economies and endanger the safety and livelihoods of people worldwide,” the space agency said.
Dagger's computer code is open source, which means it can be adopted by power grid operators, satellite controllers and telecoms companies.
Nasa said the advance warnings that Dagger provides could give companies and governments time to protect their infrastructure from an impending solar storm.
“With models like Dagger, there could one day be solar storm sirens that sound an alarm in power stations and satellite control centres around the world, just as tornado sirens wail in advance of threatening terrestrial weather in towns and cities across America,” Nasa said.
AI has been used in solar storm-predicting technology before, but this is the first time AI and real-time satellite data are being used together to make advanced predictions.