Insider threat is one of the most pressing threats in the field of information security as it leads to huge financial losses by the companies. Most of the proposed methods for detecting this threat require expensive and invasive equipment, which makes them difficult to use in practice. In this paper, we present a non-invasive method for detecting insider threat based on stress recognition using keystroke dynamics assuming that intruder experiences stress during making illegal actions, which affects the behavioral characteristics. Proposed method uses both supervised and unsupervised machine learning algorithms. As the results show, stress can provide highly valuable information for insider threat detection.