Overview
[AI Summary]: This foundational article explores how digital technologies are transforming knowledge production paradigms in humanities and social sciences through “AI for HSS” (AI for Humanities and Social Science). The authors, led by Qiu Zeqi from Peking University, argue that high-quality datafication of expert knowledge is essential for developing humanities AI, contrasting with natural sciences where mathematical expressions facilitate easier machine processing. Using legal texts as a paradigmatic example, they propose a “three-step” methodology: (1) delimiting the scope of expert knowledge that cannot be automatically extracted, (2) reconstructing knowledge structures suitable for machine processing while minimizing knowledge loss, and (3) designing technical pathways for conversion including expert-guided regular processing, simple regular processing, individualized expert knowledge import, and interference mechanisms. The article demonstrates this approach through datafying China’s Personal Information Protection Law, creating knowledge graphs that preserve legal hermeneutic knowledge while enabling machine comprehension. This work addresses the critical bottleneck preventing humanities from achieving human-machine collaborative knowledge production comparable to scientific AI developments.
- Authors: Qiu Zeqi (邱泽奇), Li Haolin (李昊林), Zhang Pingwen (张平文), Qiao Tianyu (乔天宇)
- Year: 2025
- Publication: Chinese Social Sciences (中国社会科学), Issue 7
- Institution: Peking University, UIBE, Wuhan University
- Language: Chinese