This study investigates users’ response to malfunctions in an Augmented Reality based decision support system used for sorting tasks. An initial research model and hypotheses to examine the influence of severity of malfunctions on user’s trust and intentions in a human-technology relationship were developed. A laboratory experiment was conducted in which 79 participants used Augmented Reality systems that provide incorrect instructions to varying degrees. Structural equation modeling was used to analyze the relations hypothesized in the research model. The findings indicate that the degree of technology malfunction has a significant negative influence on user’s trust, emotional state, and intentions. This study expands upon existing knowledge by integrating trust theory, psychological contract theory as well as literature on technology malfunctions to examine how users react to trust violations by emerging technologies like AR.