![]() |
|||
![]()
|
![]() |
![]() |
![]() |
3.5. TAXATIONTaxation usually involves mega-volumes of complex regulations. Taxation agencies are traditionally rated in the basement for citizen satisfaction. Expert systems are one way to handle the complexity and improve both productivity and service quality. HM Customs & Excise agency collects the value added tax (VAT) on supplies of goods and services. Businesses must comply with numerous, complex, and changing regulations or face harsh civil penalties. General auditors review client VAT systems as part of an annual audit. A U.K. auditing firm developed VAT Intelligent Assistant to help auditors appraise clients' VAT systems and procedures. In the early 1980s, the U.S. Internal Revenue Service determined that artificial intelligence techniques could be used to solve some of its information processing problems. The shear volume of information processing needed to process tax information for the individual and corporate taxpayers in a populous country was overwhelming the federal agency. The IRS began an AI project that has grown substantially over the past dozen years and now includes a number of expert systems in use (Rogers and Beckman, 1990). The IRS now uses the Correspondex expert system to generate its 15 million letter to taxpayers each year. The system not only reduced the error rate from 30% to less than 10%, but also ensures that every letter is grammatically correct and begins with a polite paragraph. The IRS's Automated Issues Identification System decides which issues should be investigated in a tax return audit. The Taxpayer Service Assistant helped the IRS provide correct tax information to taxpayers who call in their questions by phone. Other systems are in use or development. The Link Analysis System is used to identify criminal tax activity. The Reasonable Cause Determination expert system helps determine when late payment penalties should not be enforced. The International Section 482 expert system will analyze multinational corporations' income and expense calculations (Rogers and Beckman, 1990). Clearly, taxation is a domain that is well-suited for expert systems applications. Tax authorities can use expert systems to improve productivity, taxpayer satisfaction, and collections. The IRS serves as an exemplar of expert systems development and application that other government agencies should model. 4. RESEARCH ISSUESAreas for development in governmental services and operations are numerous. Some of the type of tasks that are appropriate for expert systems development are discussed below. Research questions are also identified. 4.1. TRANSITIONINGExpert systems can be developed to provide an organizational or functional memory in periods of transition. Governments, even more so than the private sector, are often characterized by periodic and frequent changes in personnel, particularly the politically elected or appointed. In addition, the diplomatic corps usually involves continuous rotation of duties from one country to another, causing diplomatic personnel to experience a learning curve each time a new post begins. Expert systems can serve as tutors to help them through the transition period more successfully and quickly. These systems will insure that the organizational knowledge is not lost with each transition or change, but is retained and made easily available for use by succeeding personnel. A number of research issues of importance are apparent. How should the required organizational intelligence be identified, captured, stored, and updated? What organizational structures are required to insure that the knowledge base is updated in a timely, effective, and cost-efficient manner? How should out-of-date knowledge be identified and purged or replaced? This last issue is particularly germane for diplomats who enter new foreign settings in which major upheavals of government have occurred. Much of the information in the knowledge base will be useless, if not dysfunctional, under those circumstances. 4.2. TRAINING AND JOB ISSUESExpert systems should be designed for use in training, identifying available jobs, forecasting skill requirements for jobs, and matching people with the appropriate jobs. As job skill requirements are identified, expert system technology can be used to train workers with the appropriate skills. Research issues to be addressed include: How is knowledge concerning job skills and needs gathered? How is such a knowledge base to be maintained properly? 4.3. TECHNICAL ASSISTANCEOne of the most onerous tasks of government and for citizens is dealing with the many rules and regulations that affect daily living. Expert systems can be used to serve as automated or semi-automated help desks. Specific applications include answering tax compliance questions, addressing compliance with health and safety regulations, identifying paperwork required to start a small business, and determining rules for obtaining permits for construction. Relevant research issues include: How can conflicts and ambiguities in laws and regulations be resolved for inclusion in a knowledge base? How are changes to regulations or laws identified and appropriate updates made to a knowledge base? How is the complexity of regulations and laws designed into a knowledge base and/or inference engine? 4.4. FRAUD, WASTE, AND ABUSEExpert systems can be developed to spot fraud, waste, and abuse caused through government contractors, inappropriate welfare recipients, or social security payments made to deceased individuals. Expert systems can also be used to examine large volumes of information in which small errors can be critical, such as criminal system records. Previous research in auditing and compliance can assist in addressing such questions as: How are likely frauds or errors identified? What data are most likely to provide evidence of errors or wrongdoing? 4.5. PLANNINGThe planning function of government is one of its largest and most complex functions. Roads, sewers, schools, and other public services require massive amounts of planning. Expert systems to assist in planning will necessarily involve large amounts of data. How is such data controlled and organized? Furthermore, how are experts needed for development of knowledge bases identified? Do experts really exist for this purpose? Or, should the expert systems development process itself be part of the expertise creation function?
|
![]() |
|
Use of this site is subject certain Terms & Conditions. Copyright (c) 1996-1999 EarthWeb, Inc.. All rights reserved. Reproduction in whole or in part in any form or medium without express written permission of EarthWeb is prohibited. Please read our privacy policy for details. |