| 英文摘要 |
This article explores the potential of generative AI as an auxiliary tool in the judicial decision-making process. It employed a literature review and indepth interviews as research methods. Based on the literature review, this article proposed a human-centered research framework that integrates aspects such as the data processing flow of generative AI and factors influencing judicial decisions. Through in-depth interviews with 2 prosecutors and 3 judges, participants interacted directly with ChatGPT on legal issues. Interview results were coded, and theoretical dialogue and refinement were performed using the conceptual framework. Findings indicate that judicial institutions favor the introduction of generative AI, particularly concerning organizational habits and personal experience limitations. However, the discretion power of prosecutors and understanding of the principle of“free proof”by judges may limit the applicability of generative AI. Finally, this article makes a theoretical contribution by proposing a conceptual framework for integrating generative AI into the judicial decision-making process. In practical suggestions, it is proposed that before adopting generative AI, judicial institutions should prioritize addressing technical and automation biases in generative AI, establish audit and review mechanisms, and develop transparent and accountable generative AI systems that can assist prosecutors and judges in improving the efficiency and quality of decision-making. |