Poster #54 - Not All Students Use AI Equally: A Multidimensional AI Usage Divide Among College Students

• In Event: Friday Poster Reception

Abstract

Background and Purpose
AI use is increasingly recognized as relevant to college students’ academic development, but students do not use AI in uniform ways, and its effects may vary by purpose of use. Prior research indicates that the adoption of AI technology may differ according to individuals' characteristics, academic program, and psychological factors, identifying these as factors that produce distinct AI usage patterns. These differences suggest the presence of an emerging “AI usage divide,” reflecting inequalities not only in access but in how AI is meaningfully integrated into students’ academic and everyday lives. Building on this perspective, the present study seeks to identify heterogeneous latent classes of AI use among Korean college students, characterize each subgroup, and examine the factors associated with class membership.

Methods 
Analyses drew on a survey administered to college students in Korea. The analytic sample (N=207) consisted of 34.8% male and 65.2% female students. 32.4% were enrolled in Nursing, 28.5% in Public Health, and 39.1% in Social Work. To determine whether it was appropriate to classify students' AI usage patterns into three latent classes—(1) passive AI users, (2) Everyday Life-Integrated AI Users, and (3) Academic & Information-Seeking AI Users—model fit indices and entropy values were examined using Latent Class Analysis. Subsequently, multinomial logistic regression analyses were conducted to identify key predictors of class membership.

Results
An examination of the relevant variables influencing the three AI usage types among Korean college students revealed that, with the Everyday Life-Integrated AI Users group as the reference group, residential area and major reached statistical significance in the Academic & Information-Seeking AI Users group. Specifically, students living in metropolitan areas (B=.815, p=.038) and Nursing majors (B=1.279, p=.007) were more likely to belong to the Academic & Information-Seeking AI Users group than those residing in small-to-medium cities or rural areas and Social Work majors, respectively. Within the Passive Users group, residential area and anxiety were statistically significant. Specifically, students living in metropolitan areas (B=1.224, p=.001) and those experiencing generalized anxiety disorder (B=1.581, p=.025) were more likely to belong to the Passive Users group than those residing in small-to-medium cities or rural areas and those with normal anxiety levels, respectively.

Conclusions
Three distinct AI usage types were identified, with group membership associated with residential area, academic major, and anxiety level. These findings suggest that AI use is shaped by structural, institutional, and psychological factors rather than preference or skill alone. Geographic disparities remained evident, disciplinary contexts influenced AI usage patterns, and anxiety was linked to passive AI use. The results support a multidimensional AI divide framework recognizing inequalities in opportunities and constraints beyond access alone. Targeted and inclusive strategies are needed to address the intersecting influences of geography, institutional context, and psychological well-being. Efforts to promote equitable AI use in higher education should reduce regional disparities, expand AI literacy and mentoring opportunities, provide discipline-specific AI training, and integrate mental health support to strengthen students’ confidence and self-efficacy in using AI tools.

Authors

• Hyuk Im, Donseo University
  Presenting Author

• Hee Lee, The University of Georgia
  Non-Presenting Co-Author

• Jinhee Koo, Pusan National University
  Non-Presenting Co-Author

• Soohong Jung, Bumin Senior Welfare Center
  Non-Presenting Co-Author

• Hansol Kim, Pusan National University
  Non-Presenting Co-Author

• JONGHYUN CHAE, Busan Geumjeong Community Senior Club
  Non-Presenting Co-Author

• Bokyung Seo, Wachi Community Welfare Center
  Non-Presenting Co-Author