인공지능 기술의 행정분야 활용에 관한 탐색적 연구 = Exploratory Study on the Use of Artificial Intelligence Technologies in Public Administration

다국어 초록 (Multilingual Abstract)

In the era of the Fourth Industrial Revolution, the field of artificial intelligence is considered the most promising and future-driven area. Many government agencies and private companies in Korea are making an effort to introduce and adopt artificial intelligence into their work. However, despite this, there are many cases in which artificial intelligence is not as effective as expected for a variety of reasons. Therefore, objective, thorough analysis and review of the introduction of artificial intelligence is essential, rather than simply holding on to a vague expectation that it will increase the efficiency and effectiveness of work. In particular, the need for such a review is much greater than in the private sector, because the projects of the administrative sector affect the lives of the public and because there is a wider range of stakeholders for any action or non-action, and the acceptability of the results demonstrated by artificial intelligence are also very important.
Therefore, this research conducts literature review, case studies, expert in-depth interviews and surveys, etc. in order to review a wide range of factors and issues necessary to understand for the adoption of artificial intelligence. We are going to propose matters in the form of a checklist that should be arranged separately into plan, design, and use stages which each research element should focus on. In this process, each research element is used to explore the complexities of the decision-making process, the level of data formatting, and the complexity of the algorithms.
Research on the introduction of artificial intelligence into the administrative sector and domestic and international case studies have identified the following issues: First, getting enough high quality data is emphasized as a key factor. Furthermore, because the social impact of artificial intelligence technology is so great, the impact of the technology should be considered. Additionally, the need for management of artificial intelligence autonomy, such as debates regarding the legal status of artificial intelligence in the future, are expected to emerge, and the need for national policies covering issues related to artificial intelligence have also been raised. Above all, the ethical problems related to artificial intelligence is a reflection of human ethics, and ethical considerations are an important factor in the process of developing artificial intelligence. In order to resolve these issues, governments and research institutes from many countries have presented principles and guidelines for developing standards for artificial intelligence systems, and can find various values such as reliability, fairness, and accountability.
For the introduction and application of an artificial intelligence system, a model should be developed as follows. Through literature review, case studies, and analysis of the standard guidelines of other governments and agencies, various factors need to be considered before embarking on artificial intelligence system development. These elements are not just listed, but have been organized into four stages of artificial intelligence system development (planning, design, pilot test, and use) through research team brainstorming, and a model is presented which arranges 24 elements into each stage. The model was developed into the final version which reflects the opinions of experts.
The academic contribution of this study is the development of a model for the administrative suitability assessment for artificial intelligence systems, presenting specific and detailed components of the model. The components of the model presented in detail are grouped according to the introduction phase of the artificial intelligence system and their association with data, algorithms, or decision making. Also, guidelines are provided for the application and dissemination of the model at administrative sites. The proliferation of the proposed model and guidelines require, among other things, innovation to establish data governance and transform into a digital-congruent government organization. In addition, training programs for cultivating a relevant workforce, the construction of infrastructure focused on cloud computing and small and medium-sized enterprises, and the establishment of laws and institutions should be carried out simultaneously.