From disengaged to self-determined: a latent transition analysis of students' AI learning motivation

Synthesis: From disengaged to self-determined — a year-long study of 2,086 secondary students across 53 schools in an AI curriculum, using Latent Transition Analysis to map students' AI learning motivation over time. Grounded in Self-Determination Theory, it identifies three motivational profiles (Disengaged, Developing, Self-Determined) at both pre- and post-test, shows most students maintaining or advancing toward higher profiles, and finds that those who reached or stayed in the Self-Determined profile showed the greatest gains in AI literacy.

Liang, Chiu, Yau, Meng, Yam, Chai & King (2026) extend Self-Determination Theory to the emerging context of AI education, moving beyond short-term AI-literacy snapshots to study how students' motivation develops across a sustained curriculum. The central message is that supporting students' psychological needs — competence, relatedness, and autonomy — drives sustained engagement and stronger AI-learning outcomes.

Method

• Design: Year-long AI curriculum with pre- and post-tests; Latent Transition Analysis to identify motivational profiles and transitions.
• Sample: 2,086 secondary students across 53 schools.
• Framework: Self-Determination Theory, conceptualizing AI learning motivation through perceived need satisfaction: competence (confidence and AI knowledge mastery), relatedness (AI for social good and ethical engagement), and autonomy (behavioral intention to continue learning AI).

Key Findings

• Three stable motivational profiles: Disengaged (low), Developing (moderate), and Self-Determined (high) need satisfaction, consistent at both time points.
• Developmental transitions: Most students maintained or moved toward higher motivational profiles over the year.
• Equity pattern: Female students and those with more prior AI learning experience were more likely to transition to higher motivational profiles.
• Motivation → AI literacy: Students who transitioned into or remained in the Self-Determined profile showed the greatest improvements in AI literacy.

What this means for practice

• Instructors. Treat Motivation as changeable and design for competence, relatedness, and autonomy throughout the year, because only the students who reached or held the Self-Determined profile showed the largest AI literacy gains.
• Instructors. Watch for downward transitions and support students who start lower — especially those with less prior AI learning experience — rather than assuming a year of AI instruction will lift everyone.
• Designers. Sustain psychological-need support across the whole curriculum instead of front-loading a short AI-literacy unit, since profiles shifted over a full year, not within a single lesson.
• Researchers. Extend SDT measurement into AI learning with behavioral indicators, because the profiles replicated in this domain and predicted learning gains but were captured only by self-report.

Limitations

• Motivation was measured only by self-report, subject to social desirability bias and students' own metacognitive limits; no behavioral or computer-log data on AI use were collected.
• Grade 9 students made up 53.2% of the sample, so the findings may not generalize to other grade levels even though grade level was included as a covariate.
• Curriculum fidelity and exact instructional methods were not systematically observed across the 53 schools, and no school-level contextual indicators (socioeconomic composition, academic performance, infrastructure) were collected.
• There was no control or comparison group, so the AI literacy gains cannot be attributed causally to the AI curriculum.

Citation

Liang, S., Yau, K. W., Meng, H., Chiu, T. K. F., Yam, Y., Chai, C. S., & King, I. (2026). From disengaged to self-determined: a latent transition analysis of students' AI learning motivation. Education and Information Technologies.