ERCIM News No.48, January 2002 [contents]
Artificial Intelligence supporting the Citizen-Friendly Information Society
by Tibor Vámos
A new project, entitled Means and systems of information society has been launched, creating an experimental but real-life environmental model of a participating democratic information society. It is a joint venture of the SZTAKI, the Budapest University of Economical Sciences and Public Administration, and the local telecommunication service provider. A three-year term is planned for the first period of the project, sponsored by a Hungarian programme for research into future trends in informatics.
The project consists of eighteen sub-tasks starting with a general survey on the present and future requirements and sociological picture of the user environment. The next item is modelling the network and its interactive dynamics. The model includes relations among the population, administration and enterprises and the services of the system. Consequences related to the organisation of the patterns form a separate sub-project. The design of the hardware and software environments is also part of the model. The influence of ideas of equal opportunity is treated in a special sub-project. Finally, real-life experimentation is envisaged. The real-life experimentation will take place in the region of Kaposvár, a relatively well-developed town in the south-western part of Hungary, with a traditional intellectual life. The objective is to surpass the usual local administration-based information society that does little more than modernise the existing administrative relations inherited from previous and very different social-political relations.
Two major sub-projects apply artificial intelligence methods in a direct way. The first is an agent-based technology for decision support, investigating the limits of automatic administrative decisions. This system is mainly rule-based, but applies case-based reasoning as well. The system should be transparent for the user-citizen, accessible not only in its progress but also clarifying the rules applied and displaying anonymous cases similar to the one concerned. In this way, the citizen may investigate his/her own chances and limitations. The flow of the decision-making process is also a check for the civilian representatives of the community and for the supervising authorities.
Earlier work in childrens care after divorce and legal decisions in bankruptcy cases are the starting points of the new efforts.
The system will be an appropriate basis for the reorganisation of the administrative agenda and the organisation as a whole.
The other artificial intelligence tool is related to language understanding and man-machine communication. The citizens should receive an interface which can be easily understood and utilised by all social strata of the population. We investigate the need for personal communication and the limits of machine interfaces, bearing in mind that the problem is educational and psychological rather than technological. We intend to surpass the rigidity of the usual questionnaires and try to apply machine understanding of natural language communication, albeit limited to the subject at hand.
Understanding in general is a key problem in artificial intelligence. In our case it should be a man-machine system applying the positive qualities of the machine objectivity, working transparency and of the human face of the helpful agent. Hungarian presents a special task for achieving understanding, since its structure differs so markedly from Indo-European languages. The project involves the usual problems of contradictions between privacy and control action, and their legal, sociological, educational and technological issues. Democracy is exercised using an electronic agora, and involves all the problems of democracy and efficient work, civil participation and professional administration, majority opinions vs. minority rights, free access and authority.
Beyond the regular administrative tasks, some other public services will be included in the system. Citizens who do not possess direct access in their homes are and will continue to be served by so-called telehouses, a public access network available through schools, libraries, post offices and cafes.
Several similar initiatives have been launched, mostly in European countries and in Japan, and a special study has analysed the lessons that emerged. However, this project differs significantly in two respects. The first is the local environment, its sociological and cultural background. Hungary has less experience with democracy, especially in the conscious participation of society in the common agenda. The system should therefore focus on these issues. The other difference is the more advanced level of application of artificial intelligence tools, as detailed above. We hope that after the design of the relevant models, ie, after about a year, we can join the other European efforts. The fifth and the future sixth framework of the European Unions research and development agenda have a special emphasis on systems like ours.