The study presented in this paper shows that service users can have low confidence in a service provider's ability to protect their personal information even if those service users trust the overall brand. Today, on-line services are not specifically designed to promote a service user's confidence building. As a result, service users have to depend on off-line techniques to build confidence in their information practices. One implication of not having effective support for confidence building designed into the on-line service is that, despite costly investment in trust marks, security technologies and brand development, service users will continue to give false information, limit the extent of their engagement in on-line services and avoid registration with on-line services. In the era of on-line public services delivery, this pattern of privacy protection practice potentially has devastating consequences for public service delivery and the ability of the most vulnerable to receive the public service support that they need. The study also indicates that providing interaction possibilities through social computing as part of the service design is one way to help build service user confidence. This paper concludes with examples of social computing used for this purpose.