An ODeL-based teacher professional development framework for teacher-facilitated human–AI collaboration in primary mathematics problem-solving: a conceptual analysis

Source abstract

Artificial intelligence (AI) tools are increasingly present in primary mathematics classrooms, yet their educational value is widely argued to depend on how teachers use them. Existing models explain parts of this problem: AI-TPACK describes the knowledge teachers need, teacher-in-the-loop and human-centred AI models describe how control is shared between teachers and systems, and orchestration research describes how teachers coordinate classroom activity. What remains under-theorised is the pathway linking the way open distance e-learning (ODeL) professional development builds teachers’ capacity to its enactment when AI scaffolds learners’ problem-solving. This conceptual article asks: through which mechanisms, and under which conditions, can ODeL-based professional development be expected to enable teacher-facilitated human–AI collaboration that supports primary learners’ mathematical problem-solving? The analysis combined an integrative review of peer-reviewed and selected grey literature with theory synthesis: sources were identified through structured database searches, screened against pre-specified criteria, coded for constructs, mechanisms and boundary conditions, and organised by level of analysis. The primary context is South African primary schooling (Grades 1–7) and low-bandwidth, predominantly asynchronous ODeL. Positioning four perspectives within a multilevel meta-model, Intelligent-TPACK at the teacher level, Zone of Proximal Development scaffolding and social constructivism at the interaction and classroom levels, and Polya’s heuristic as an analytic lens at the learner level, with transactional distance theory at the programme level, I propose a three-layer framework that distinguishes inputs, mechanisms, practices and outcomes, and derive seven testable propositions. These include teacher orchestration as a mediator between AI-TPACK and scaffolding quality; infrastructure readiness, learner age, language, and AI type as moderators; and governance and children’s rights as cross-cutting boundary conditions. Unlike models that specify teacher knowledge or the division of control alone, the framework specifies the professional development pathway through which human–AI collaboration is expected to benefit learners. I propose that well-designed ODeL professional development is an enabling condition for such benefit, not a guarantee of it, and close with implementation and evaluation indicators for teacher educators, policymakers, school leaders, researchers and learners.