复旦大学国家智能评价与治理实验基地硕士研究生钱鹏博与复旦大学新闻学院谢兴政副研究员在人机交互领域知名Q1期刊（SSCI&SCI双收录）《国际人机交互杂志》（International Journal of Human-Computer Interaction）合作发表《任务技术匹配导致冲突：生成式人工智能对人文社科研究科学创造力的双刃剑效应》（Task–Technology Fit Leads to Conflict: The Double-Edged-Sword Effect of Generative Artificial Intelligence on Scientific Creative Performance in Humanities and Social Sciences Research）研究论文。该研究旨在揭示生成式人工智能在人文社科研究中影响学者科学创造力的复杂机制。

摘要：This study examines the double-edged sword effect of task-technology fit (TTF) on scientific creative performance in the humanities and social sciences, utilizing a mixed-methods approach. Analyzing data from 405 Chinese HSS scholars through structural equation modeling and conducting thematic analysis with 12 in-depth interviews revealed that TTF generates a paradox by enhancing AI literacy while increasing AI dependence. AI literacy promotes augmentation interactions, boosting scientific creativity, whereas dependence leads to automation patterns that hinder it. Demographic variations show female and senior scholars exhibit higher levels of dependence, with AI usage frequency demonstrating a non-linear relationship with dependence. Qualitative insights highlight distinct methodological orientations toward GAI among subdisciplines, raising concerns about ethical implications, dependency, and academic inequality. This research challenges the assumption that technology-task fit always yields positive outcomes and introduces the “AI Empowerment-Inhibition Paradox” as a framework for understanding these dynamics.

关键词：Task-technology fit; scientific creative performance; double-edged-sword effect; generative AI

来源期刊：International Journal of Human-Computer Interaction

发表时间：2025-07-14

DOI：https://doi.org/10.1080/10447318.2025.2526580

收录数据库：Web of Science