A practical method, based on the Akaike Information Criterion (AIC), is used to detect automatically the first P-wave arrival times during Acoustic Emissions (AE) monitoring. This method is used to increase the accuracy of detection procedure and to reject the data errors automatically. Two parameters, the quantification of the degree of certainty and the apparent velocity between observation sensors, were proposed to check the validity of AE events and to neglect errors. Both the original AIC-picker and improved AIC-picker algorithms are applied to AE signal processing. It can be shown that the improved AIC-picker is a reliable tool for automatic onset detection of AE signals.