Teacher competence and institutional governance for learner data privacy in AI mediated education

Teachers can recognise a privacy problem yet lack the authority or resources needed to act on that judgement. Conversely, institutions can provide technically safeguarded services while teachers disclose unnecessary information or accept unsupported inferences. This Hypothesis and Theory article develops the Learner Data Protection Competence and Governance Framework to explain these two failures. A purposive conceptual synthesis connects contextual integrity with distinctions between demonstrated competence, confidence, perceived control and actual institutional provision. The framework relates risk recognition, defensible decisions, institutional enablement and accountable review. Its central prediction distinguishes support that complements professional judgement in discretionary tasks from safeguards that substitute for particular skills in constrained tasks. A second central prediction separates access to reporting, implementation of corrective action and subsequent recurrence. Three supporting propositions address recognition, confidence calibration and purpose-specific authorisation. Each proposition specifies separate measures, rival explanations and conditions for disconfirmation. Two hypothetical cases illustrate competing educational and privacy considerations and show why institutional approval is insufficient to establish ethical justification. A bounded application to teacher education and school placements in Zambia identifies context-informed predictions concerning divided authority, advice latency, language and connectivity. The framework offers an explanatory model with normative criteria and assessment implications. Neither the reviewed evidence nor the worked cases validate its causal relationships or demonstrate reduced privacy harm; the contribution is a more discriminating and testable account of how professional judgement interacts with institutional arrangements.