AbstractEarthquake early warning studies are shifting real‐time seismology in earthquake science. They provide methods to rapidly assess earthquakes to predict damaging ground shaking. Preventing false alarms from these systems is key. Here we developed a simple, robust algorithm, Authorizing GRound shaking for Earthquake Early warning Systems (AGREEs), to reduce falsely issued alarms. This is a network threshold‐based algorithm, which differs from existing approaches based on apparent velocity of P and S waves. AGREEs is designed to function as an external module to support existing earthquake early warning systems (EEWSs) and filters out the false events, by evaluating actual shaking near the epicenter. Our retrospective analyses of the 2009 L'Aquila and 2012 Emilia earthquakes show that AGREEs could help an EEWS by confirming the epicentral intensity. Furthermore, AGREEs is able to effectively identify three false events due to a storm, a teleseismic earthquake, and broken sensors in Irpinia Seismic Network, Italy.