Facial expressions are one of the most salient cues of affect. However, they are also a communication tool that can be expressed voluntarily. Spontaneous and posed facial expressions have different characteristics. In this study, we show the potential of using Electromyography (EMG) in a wearable device to automatically differentiate between posed and spontaneous smiles. This classification is moderately successful when using spatial and magnitude features, and increases significantly if temporal features are included. Hence, the high temporal resolution of EMG-based detection is advantageous in this task.