공법상 가치와 인공지능 시스템 — 공적 영역에서의 인공지능 활용을 중심으로 —

초록

The use of artificial intelligence in the public domain needs to be subject to certain control by public laws. Unlike the private sector, there is less resistance to direct regulation, so it is also appropriate as a test bed for artificial intelligence regulation. In this study, based on the principle of public law, preliminary control of the action itself or a method to return the effect of the action was sought. As a first step, cases in which efforts were made to realize certain values for the use of artificial intelligence systems are explored. Although each subject has different perspectives and contents because they have different interests, the international community has generally drawn common artificial intelligence ethics principles over the past few years. In the second step, elements applicable to the use of artificial intelligence systems were extracted from various values or concepts pursued by legal order. Public law values such as human dignity and worth and the right to pursuit of happines, the principles of equality, personal safety and freedom, protection of privacy and socio-economic safety and freedom are also meaningful in the use of artificial intelligence systems. Finally, in consideration of the driving method and characteristics of the artificial intelligence system, methods were devised to realize the public law principles differently from the existing actions under public law. A method of mandatory control that declares the basic principles of public law legally or through internal administrative rules can be considered. In the preliminary stage, an impact assessment that evaluates the risk of using artificial intelligence systems can be used. The entire social community may participate in the discussion to form a governance that constitutes discourse. For AI systems that do not conform to the values of the public law, it is also possible to consider excluding use through a certain procedural control. Additional considerations are also required to effectively apply these methods. It is necessary to think about how to derive the concept of values under public law that can be used in the application of artificial intelligence systems from embodied and substantiated values. Subsequently, it should be recognized that the means to secure public law values do not exert utility with a single means, but are more effective when multiple means are used at the same time.

키워드