Profiles of AI Dependency: A Latent Class Analysis of Filipino Students' Academic Competencies

Abstract: The increasing dependency among Filipino college students on AI poses concerns about the potential decline of fundamental academic competencies. This study examines the extent of AI dependency and its perceived effects on students' critical thinking, writing skills, learning independence, research skills, and academic engagement. Using a cross-sectional research design, data was collected from 651 students enrolled in higher education institutions (HEIs) in Pampanga, Philippines accredited by the Commission on Higher Education. The survey data was analyzed using Latent Class Analysis (LCA) to identify AI dependency patterns. Findings indicated that students show moderate to high AI dependency, specifically in research and writing tasks. LCA identified four distinct profiles: highly engaged independent learners, selective AI users, moderate AI users, and AI-dependent learners. Notably, AI-dependent learners demonstrated the weakest academic competencies, with significant dependency on AI-generated outputs. The study highlights the need to foster educational policies that integrate AI literacy while preserving essential academic skills. HEIs must also balance technological advancements with curriculum adaptations to promote critical thinking and ethical use of AI. Future research may explore the longitudinal impacts and intervention strategies to mitigate academic skill erosion caused by AI dependency.

• The paper identifies four distinct AI dependency profiles using latent class analysis on data from 651 Filipino college students.
• It employs reliable survey methods with high Cronbach's alpha scores (0.92–0.97) and various statistical techniques to link AI usage with key academic competencies.
• Findings indicate that increased AI reliance is correlated with reduced critical thinking and learning autonomy, urging urgent curricular reforms.

Latent Class Profiling of AI Dependency and Academic Competency Erosion in Filipino Higher Education

Research Objectives and Context

The study "Profiles of AI Dependency: A Latent Class Analysis of Filipino Students' Academic Competencies" (2604.27349) examines the extent and ramifications of AI dependency among college students in Pampanga, Philippines. The research is motivated by the rapid proliferation of generative AI tools in academic settings and the resulting concerns about cognitive offloading, the erosion of traditional academic skills, and shifting perceptions of authorship and learning authenticity. The study specifically assesses the relationship between AI reliance and five key competencies: critical thinking, writing skills, learning independence, research skills, and academic engagement.

By leveraging Latent Class Analysis (LCA), the study aims to derive behavioral profiles of AI usage among Filipino tertiary students and to relate these profiles to academic performance and demographic variables. This approach addresses gaps in the literature that often neglect person-centered profiling and the regional specificity of AI’s educational impact, particularly in the context of Southeast Asia.

Methodology

A cross-sectional design was employed, surveying 651 students from Commission on Higher Education-accredited institutions across Pampanga, a representative region for technologically progressive Philippine higher education. The research instrument, developed in accordance with current literature, exhibited high reliability with Cronbach's alpha values ranging from 0.92 to 0.97 across the five competency domains.

Data collection was carried out using Google Forms over four months, utilizing snowball sampling to reach students with generative AI experience. Analyses included descriptive statistics, LCA for profile identification (model fit determined by AIC and BIC), and Chi-square tests for demographic associations.

Empirical Findings

Patterns of AI Dependency

Descriptive statistics indicated a moderate to high frequency of AI tool usage for research and writing-related tasks (notably, uncritical acceptance of AI outputs: $M=3.23$, research suggestions: $M=3.13$). Dependencies were lower for academic decision-making and collaborative tasks, but dependency was found to be systematic, not sporadic, with general mean scores in the range $M=2.72$ to $M=2.90$ for core competencies—indicative of pervasive integration of AI as a substitute for authentic academic engagement.

Latent Class Profiles

LCA yielded an optimal four-class solution:

• Highly Engaged Independent Learners: Demonstrated strongest competencies and minimal AI dependency, adopting AI selectively.
• Selective AI Users: Maintained strong research and writing skills and high engagement, using AI as a complementary tool rather than a crutch.
• Moderate AI Users: Displayed balanced skills with signs of dependency, especially diminished learning independence, suggesting emerging substitutive tendencies.
• AI-Dependent Learners: Exhibited the weakest functional competencies, with pronounced reliance on AI for writing, research, and problem-solving, and significantly reduced learning autonomy.

These profiles underscore a spectrum of AI integration, from augmentation to functional substitution and skill stagnation.

Demographic Analyses

No significant gender or academic performance associations with AI dependence were found (p > 0.05), counter to some prior literature suggesting differential adoption patterns by gender. However, a significant effect was observed for academic year (p = 0.003), with senior students exhibiting stronger dependency, attributed plausibly to escalating academic workload and greater cumulative exposure to AI tools. This suggests an accumulative risk of dependency as students advance, regardless of baseline academic ability.

Theoretical and Practical Implications

The findings raise concerns about the long-term trajectory of academic competency in the presence of generative AI, especially for learners classified as AI-dependent. If current trends persist, widespread cognitive offloading may result in atrophied critical thinking, writing, and autonomous research skills, particularly as AI systems become further embedded in academic workflows.

From the institutional perspective, these results necessitate urgent curricular and policy innovations. There is a critical need for AI literacy programs that go beyond technical proficiency by fostering epistemic vigilance, information validation, and self-regulatory strategies. Interventions should be tailored: for instance, targeting Moderate AI Users with training in analytical verification, supporting AI-Dependent Learners with remedial skill-building, and deploying Highly Engaged Independent Learners as peer mentors. At the systemic level, educational ecosystems must be recalibrated to leverage AI augmentatively rather than substitutively.

The study also sets the stage for longitudinal research to track the dynamic evolution of these dependency profiles and to evaluate the impact of policy interventions on skill retention and academic authenticity.

Conclusion

This research provides a nuanced, empirical mapping of AI dependency among Filipino higher education students, identifying distinct behavioral phenotypes and their association with academic skill erosion. The absence of significant gender or performance-based differentiation, alongside the pronounced effect of academic progression, signals a systemic risk of dependency-driven academic skill decline. The necessity for purposeful educational reforms and the embedding of robust AI literacy into curriculum policy are underscored as imperatives for preserving foundational competencies amidst technological transformation.