A Study on the Framework for Evaluating the Ethics and Trustworthiness of Generative AI: A Case of Generative AI Chatbot Services

This study proposes a comprehensive framework to analyze and evaluate the ethical and trustworthiness problems of Generative AI. Generative AI such as ChatGPT has innovative potential but simultaneously causes ethical and social issues including bias, harmfulness, copyright infringement, privacy infringement, and hallucinations. Existing artificial intelligence (AI) evaluation methodologies mainly focus on performance and accuracy, limiting their ability to address these complex problems. This study emphasizes the need for new evaluation criteria that are human-centered and consider social impact. This study defines key factors for evaluating the ethics and reliability of Generative AI, including fairness, transparency, accountability, safety, privacy, accuracy, consistency, robustness, explainability, copyright protection, and source traceability, presenting detailed indicators and evaluation methodologies for each factor. AI ethics policies and guidelines of major countries such as South Korea, the United States, the European Union, and China are compared and analyzed to derive implications for each country. The proposed evaluation framework is applicable throughout the AI system life cycle and contributes to effective identification and management of AI ethics and reliability issues by integrating multidisciplinary perspectives with technical evaluation. To demonstrate practical applicability, pilot experiments in a Generative AI chatbot system verify the usefulness of the indicators, laying an academic foundation for responsible AI development. The framework provides practical guidelines for stakeholders such as policymakers, developers, and users, contributing to the positive social impact of AI technology. Received: 29 August 2025 | Revised: 12 January 2026 | Accepted: 17 March 2026 Conflicts of Interest The authors declare that they have no conflicts of interest to this work. Data Availability Statement Data sharing is not applicable to this article as no new data were created or analyzed in this study. Author Contribution Statement Cheonsu Jeong: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Resources, Data curation, Writing – original draft, Writing – review & editing, Visualization, Supervision, Project administration. Seunghyun Lee: Software, Validation, Formal analysis, Investigation, Resources, Data curation, Writing – review & editing, Visualization. Seonhee Jeong: Software, Validation, Formal analysis, Investigation, Resources, Data curation, Writing – review & editing, Visualization. Sungsu Kim: Software, Validation, Formal analysis, Investigation, Resources, Data curation, Writing – review & editing, Visualization.