XU Xiaojing, WU Zhaotong, SUN Qiao, YAN Hui
LIBRARY TRIBUNE.
2026, 46(7):
90-104.
Based on the affordance actualization theory,this paper analyzes the data collected from researchers through in-depth interviews and app walkthroughs of GenAI-assisted reading tools,employing a three-level coding scheme,and proposes a model of the GenAI-enabled academic reading mechanism. The study indicates that:Information overload,cognitive barriers,communication constraints constitute major obstacles in academic reading. The text-processing potential of GenAI,including information reconstruction and content parsing,provides technical support for breaking these bottlenecks;By aligning technical characteristics with contextual needs,large language models can serve not only as cognitive aids that help overcome barriers to knowledge acquisition,but also,through their communicative and interactive capabilities,as unique collaborative partners in the reading process,demonstrating dual affordances of transcendence and interactivity;In academic reading contexts,GenAI’s affordance manifests as cognitive outsourcing,encompassing three scenarios (screening and skimming,in-depth reading,and extended application) as well as three modes (tool-dependent,assistive-enhanced,and cautiously reserved). This may bring about cognitive optimization effects,but also carries the risk of cognitive dependence. In terms of process,the study examines how GenAI influences academic reading workflows such as literature screening and problem-solving;in terms of outcome,it identifies the efficiency gains brought by GenAI and the differences in evaluation between novice and experienced researchers.