개인정보가 포함된 인공지능 학습데이터의 처리 근거 - 공개된 개인정보, 이미 보유한 이용자 개인정보를 중심으로 -

Since the emergence of ChatGPT in 2022, generative artificial intelligence has brought rapid changes across industries and daily life. Because artificial intelligence model performance depends on the quality and scale of training data, companies are building large-scale training datasets by utilizing publicly available personal information and personal information of users they already possess. However, legal uncertainty persists as it remains unclear whether the processing of artificial intelligence training data containing personal information is lawful under personal information protection laws. Accordingly, this paper examines the concept, characteristics, and privacy risks of artificial intelligence training, and reviews interpretations regarding the processing of artificial intelligence training data, focusing on publicly available personal information and personal information of existing users, while analyzing their limitations. Ultimately, to achieve a balance between personal information protection and artificial intelligence industry development, it is urgent to clarify the lawful basis for processing artificial intelligence training data containing personal information and establish legal standards that can guarantee the rights of data subjects. Through this, it will be possible to resolve legal uncertainty for businesses, protect the right to self-determination of personal information, and promote trust-based development of the artificial intelligence industry.