Teacher education for artificial intelligence literacy through a self-determination theory perspective

Synthesis: Chiu, Bali, Tondeur, Howard, and Chan (2026) apply self-determination theory (SDT) to investigate how need-supportive professional development (PD) impacts teachers' AI literacy, attitudes, anxiety, and engagement in online professional learning communities (PLCs). Using a sequential mixed-methods approach with 382 secondary school teachers, they found that need-supportive PD enhances AI literacy and fosters sustained behavioral engagement in PLCs, with AI literacy emerging as the key cognitive driver of sustained participation. Qualitative analysis identified nine specific design strategies for PD that satisfy teachers' needs for autonomy, competence, and relatedness — bridging the often-overlooked connection between isolated PD and PLCs to support the scaling and Sustainability of teacher AI learning.

Key Findings

• Need-supportive PD positively predicts needs satisfaction (β = .77, p < .001) and behavioral engagement in PLCs (β = .47, p < .001) — a well-designed PD can spark ongoing, Collaborative Learning.
• Needs satisfaction significantly predicts AI literacy (β = .52), attitude (β = .68), and reduces AI anxiety (β = −.66) — PD content alone is insufficient; how it is delivered (whether it supports psychological needs) determines its effectiveness in changing teachers' AI knowledge, feelings, and beliefs.
• AI literacy is the key cognitive gateway to sustained engagement: among the AI learning outcomes, only AI literacy had a significant direct effect on behavioral engagement in PLCs (β = .39, p < .001); attitude (p = .72) and anxiety (p = .18) were non-significant — teachers need a foundation of ability before actively engaging in peer-to-peer learning.
• The theoretical contribution is a validated motivational pathway: PD needs support → needs satisfaction → cognitive (AI literacy) and affective (attitude, anxiety) learning → sustained behavioral engagement, with AI literacy as the primary cognitive driver.
• Nine design strategies were identified: four autonomy-supportive (elective AI topic selection, collaborative AI resource development/co-design, individualized AI integration targets, non-mandatory task completion), two competence-supportive (expert-peer feedback on AI learning designs, public sharing of AI teaching artifacts), and three relatedness-supportive (cross-disciplinary AI design teams, mentorship, mission-alignment of AI learning).

Study Design & Method

This sequential mixed-methods study examined 382 secondary school teachers engaged in AI-focused professional development. Quantitative phase: participants completed validated SDT-based questionnaires (perceived PD support, autonomy, relatedness, competence) plus measures of AI attitude, AI anxiety, and AI literacy (a 30-item multiple-choice test drawn from the AI4future item bank), and behavioral engagement in online PLCs was objectively measured by posts, replies, reads, sharing, and collaborations. Data were analyzed with confirmatory factor analysis (CFA) and structural equation modeling (SEM), with good model fit (χ²/df = 1.43, RMSEA = .03, CFI = .99). Qualitative phase: deductive content analysis of interviews using SDT as the analytic framework identified the nine design strategies, with two independent raters and a moderator (inter-rater reliability = 0.88).

What this means for practice

• Faculty developers. Build the nine need-supportive strategies into the structure of the program (elective AI topics, collaborative co-design, individualized integration targets, non-mandatory completion, expert peer feedback, cross-disciplinary teams, mentorship, mission alignment) rather than adding more AI content to a transmission-style workshop.
• Faculty developers. Sequence foundational AI Literacy before expecting peer participation: AI literacy was the only learning outcome with a significant direct effect on behavioral engagement in the PLC (β = .39), while attitude (p = .72) and anxiety (p = .18) were not.
• Faculty developers. Treat the program as a catalyst and plan the handoff that keeps the PLC need-supportive after the structured sessions end, since sustaining that culture is where the design is most exposed.
• Administrators. Remove the systemic constraints that undercut need-supportive PD — rigid curricula, limited resources, top-down mandates — because needs satisfaction is what carried the effects on AI literacy (β = .52), attitude (β = .68), and reduced anxiety (β = .66).

Limitations

The study's context is secondary school teachers (largely in Chinese/Hong Kong and East Asian settings), bounding generalizability to other regions and K-12 contexts. The authors acknowledge that sustaining a need-supportive culture within PLCs after the structured program ends is challenging, that systemic constraints (rigid curricula, limited resources, top-down mandates) can undermine need-supportive PD, and that the strategies may overlook the mediating role of contextual factors like teacher prior AI experience or entrenched school culture. The AI literacy test was a specific objective measure developed for the AI4future project.

Citation

Chiu, T. K. F., Bali, S., Tondeur, J., Howard, S., & Chan, K. K. H. (2026). Teacher education for artificial intelligence literacy through a self-determination theory perspective. European Journal of Teacher Education.