공개된 개인정보의 법적 취급에 대한 검토 ― AI학습용 데이터로서 활용방안을 중심으로 ―

Utilization of publicly available personal information is an extremely important and sensitive issue in the age of artificial intelligence and big data. However, in Korea, there are currently no special provisions in the laws regarding this, and the legal standards suggested by precedents are also ambiguous. Therefore, this study tried to find a way to make it legal to process publicly available personal information as data for artificial intelligence learning. In particular, it was attempted to explore the nature of the rights of data subjects and to find ways not to violate them. The results of the study are as follows.: First, concerns about personal information being processed as data for AI learning are due to the possibility of identifying a specific person or including the original data in an algorithm that is a learning result. Then, if processing for identification of a specific person is prohibited and the inclusion of original data in the result value of the AI algorithm is fundamentally prohibited, processing of publicly available personal information as data for AI learning should be allowed. Second, as a way to legalize ‘publicly available personal information’ so that it can be used for AI learning, the following is proposed. i) A method of excluding ‘publicly available personal information’ from the concept of ‘personal information’ to which the Personal Information Protection Act applies ii) A method to make a legal basis for freely using publicly available personal information as data for artificial intelligence learning unless it is to identify a specific person, and iii) A method to specify standards and measures for legally using publicly available personal information through guidelines, etc. to clarify the contents of ‘objective judgment’ according to current precedents.