Personalized Vehicle Insurance Rates - A Case for Client-Side Personalization in Ubiquitous Computing Vlad Coroama, Marc Langheinrich Abstract ======== In the foreseeable future, miniature sensor and networking technology will allow insurers to offer personalized insurance rates based not only on demographic data, but on actual customer behavior. Behavior-based personalized rates can significantly lessen the traditional information asymmetry between insurers and insured, and thus prompt significant cost savings both on a personal and societal level. However, initial prototypes have so far required consumers to disclose their entire data set in order to enjoy personalized cost savings. We have opted to follow a more privacy-friendly approach, by performing the entire personalization on a trusted platform on the client-side and only reporting gross rates to the insurer. This article describes our system and a set of requirements we derived for client-side personalized insurance rates.