Information Valuation Policies for Explainable Trustworthiness Assessment in E-Services

Abstract

Information e-services are useful for exchanging information among many users, whether human or automated agent; however, e-service users are susceptible to risk of inaccurate information, since users have no traditional face-to-face interaction or past relational history with information providers. To encourage use of information e-services, individuals must have technology to assess information accuracy and information source trustworthiness. This research permits automated e-service users—here called agents—acting on behalf of humans, to employ policies, or heuristics, for predicting information accuracy when numerous pieces of information are received from multiple sources. These intuitive policies draw from human strategies for evaluating the trustworthiness of information to not only identify accurate information, but also distinguish untrustworthy information providers. These policies allow the agent to build a user’s confidence in the trust assessment technology by creating justifications for trust assessment decisions and identifying particular policies integral to a given assessment decision.