AI Predicts Earthquakes Days in Advance, Say Researchers
In a groundbreaking development, scientists have harnessed artificial intelligence (AI) to predict earthquakes several days before they occur, offering a potential leap forward in disaster preparedness and risk mitigation.
A notable study by researchers at the University of Texas at Austin demonstrated the efficacy of an AI algorithm trained on five years of seismic data from China's Sichuan and Yunnan provinces. During a seven-month trial, this AI system accurately forecasted 14 out of 15 significant earthquakes up to a week in advance, achieving a 70% success rate. The predictions were within approximately 200 miles of the actual epicenters and closely matched the observed magnitudes.The AI model operates by analyzing continuous seismic recordings to detect subtle patterns and anomalies that often precede major seismic events. By identifying these precursors, the system can generate alerts days before an earthquake strikes, providing valuable time for emergency response and public safety measures. However, the technology is not without limitations. The same study reported eight false positives—instances where the AI predicted earthquakes that did not occur—and one missed event. These findings highlight the need for further refinement to enhance the model's precision and reduce false alarms.