With the increasing diversity and complexity of the public’s problems and needs, expectations for the use of AI are growing. AI has the potential to solve many social problems, but AI systems with complex logic are often referred to as “black boxes,” and their decision-making processes are generally difficult for humans to understand. This difficulty of interpretation can lead to risks such as unpredictability of AI responses to environmental changes, lack of impartiality, and opacity of accountability. Particularly in AI operations that have a significant impact on society and individuals, these issues can create complex situations that raise ethical and legal problems. In addition, the uncertainty inherent in AI models makes it difficult to adequately address these risks.