Real-time recognition of Activities of Daily Living (ADLs) enables surveillance, security, and health monitoring applications. Establishing systems to recognize ADLs non-intrusively and passively, would facilitate smart spaces by providing assistive living for seniors and people with disabilities. This work applies artificial intelligence to pervasive Wi-Fi technology to measure motion and distinguish the patterns associated with dynamic activities. Processed signals are segmented into labelled sequences. Statistical features are extracted to enable activity recognition. Classifiers are tested to predict activities. Support vector machines achieve the best performance at distinguishing still, sitting-down, and standing-up activity classes. Results on data collected by one subject achieve a classification accuracy of 98.8% for sitting and standing activities at one room location, within one experimental setup, and 98.5% for various activity locations, within different experimental setups. Results on data collected by four subjects achieve an accuracy of 97.6% for sitting and standing activities at random room locations.