Incorporating moral motivation in automatic summary generation of literary fiction

Automatic summary generation aims to condense knowledge and improve users' information retrieval and learning efficiency across various fields. In digital reading, the goal of attracting users' attention and guiding in-depth reading has led to a change in the function of summarization. The moralizing potential of literary fiction represents a unique feature that draws readers' attention and facilitates the retention of moral information. Yet, prior research has predominantly emphasized information summarization, thereby neglecting these underlying moral-cognitive mechanisms. This study introduces the concept of psychological moral motivation and constructs the MoralTextRank model, using 50 literary fiction works to generate summaries containing moral information. Evaluation indexes were designed, and tests were conducted with 120 participants to assess reading effects. The results show that summaries containing moral information significantly attract users' reading attention compared to both neutral and low-morality summaries, particularly among male users ( t l o w = − 3.03 , p = 0.0034 ; t h i g h = − 2.9 , p = 0.0049 ). Specifically, male participants showed 8.2 % and 15.2 % more engagement with high-morality summaries than with low- and no-morality ones, respectively. This paper argues that domain-specific needs significantly influence the purpose and design of summary generation. Integrating moral information into literary fiction summaries can effectively capture readers’ attention and enrich digital reading experiences. In turn, this practice can enhance the efficiency of attention allocation within the digital content environment. This research aims to optimize information technology design and processes by integrating socio-cultural factors, thereby enriching its socio-cultural connotations. • Moral information content enhances users' attention and supports content dissemination. • Summaries emphasizing moral narratives promote deeper literary-fiction engagement. • Embedding moral elements in summaries strengthens reading motivation and educational value. • Integrating moral feedback into information technology enriches socio-humanistic value of content.