Mobile computing and communication devices are open to surreptitious privacy attacks using emotion detection techniques; largely utilising work carried out in the area of voice stress analysis (VSA). This paper extends some work in the area of removing emotion cues in the voice, specifically focusing on lie detection and presents the results of a pilot study indicating that the use of mobile phones in situations of stress is common and that awareness of VSA is low. Existing strategies for the removal or modification of emotion cues, based on models of synthesis are considered and weaknesses are identified.