Laughter is a naturally occurring feature in speech and social interactions. Human intelligence can identify people by their laughter, but this has not been explored as a potential biometric in person identification systems. This study proposes a novel behavioural biometric based on individual laughter signatures. Mel frequency cepstral coefficients (MFCC) features were extracted and Kruskal-Wallis test was performed on each coefficient. A dynamic-average Mel frequency cepstral coefficients (DA-MFCC) was developed from the typical MFCC features for system training using Gaussian mixture model (GMM) and support vector machine (SVM). Test results showed an accuracy of 90%-person identification for SVM while the GMM was 65%. The use of GA-MFCC improved the accuracy of the system by 5.06% and 2.99% for GMM and SVM respectively. Laughter has thus been shown to be a viable biometric feature for person identification which can be embedded into artificial intelligence systems in diverse applications.