| Abstract | With the rapid employment of the Internet and e-commerce in business sectors, e-government research has emerged with the objective of bringing about a better model of government services and interactions with citizens. To model effective e-government services based on citizen expectations, this dissertation asserts that first, the design of e-government service must encompass good governance principles as an objective and a priority, and second, it is necessary to take into account the unique nature of the government, namely the rule of law, in order to understand it's implications in supporting or hindering good governance outcomes. This dissertation specifically examines the technical aspects of e-government services and applications under the rule of law that influence the design of e-government applications. The research explores two types of legal language that govern e-government applications. The first is a rule-based type of legal language. This type of law is explicit, clear, and strict. The use of rule-based logic is very common in the design of e-government applications; however, the complexity and myriad nature of legal rules are major problems in the design and decision-making. The second is an "open-textured" legal language, where there is an absence of previously fixed, relatively clear, and binding legal standards. In this case, laws leave room for interpretation for an official who must exercise discretion in a non-arbitrary manner. In dealing with rule-based e-goveriunent, this dissertation investigates a case study of the Rules of Origin verification services. This dissertation explores the implications of Artificial Intelligence (AI) and laws in the design of e-government applications. This research suggests that the Composite Act Frame technique extended from van Kralingen and Visser's legal frame-based ontology deal with the complexity and voluminous nature of the Rules of Origin criteria. To prove the generic and extendibility aspect of ttJ-e Rules of Origin knowledge representation model, an assessment test with different criteria for the Rules of Origin is conducted. The implementation of the Rules of Origin knowledge representations to support the web-based e-government services is accomplished through the system called "Rules of Origin Verification Systems: ROVERs". The Rules of Origin knowledge representation in this dissertation represents a generic model that may be applied and extended to other agreements for modeling purposes. With respect to the "open-textured" language of law or the discretionary-based e¬government, this dissertation examines the concept of administrative discretion in relation to artificial intelligence and laws. An 'open government' approach is adopted as a means to efficaciously prevent whim, caprice, chance, or routine bureaucratic discretionary decisions. To achieve transparency and openness in administrative discretionary decision-making, i.e. adjudication, modified adjudication, and policy issue, the citizen consultation process is presented. The E-government Discretionary Framework is proposed as a lens with which to analyze and design e-govemment applications. A Citizen Consultation Model is presented as a guideline for promoting transparency from the modified adjudication decision-making. The prototype of the Citizen Consultation System is developed and the evaluation results are illustrated |