Designing intentional AI-augmented learning: social learning theory extended through the human-machine-human model

Abstract

Drawing on social learning theory (SLT), this paper aims to introduce the human-machine-human (HMH) model as a diagnostic framework for identifying structural gaps that emerge when artificial intelligence (AI) mediates workplace learning. It enables organizations to assess their readiness to deploy AI-augmented learning systems intentionally. The paper builds on Bandura’s SLT and recent literature to conceptualize three structural gaps (trust, reinforcement and tacit knowledge) arising from AI-mediated learning management systems. These gaps arise because algorithmic mediation cannot fully replicate the relational authority, tacit knowledge and emotional tuning inherent in human-to-human interactions. It identifies four conditions and diagnostic questions, centered around digital fluency, self-regulation, psychological safety and AI governance, that must be addressed for a successful implementation that yields meaningful results for both learners and organizations. Empirical validation of the HMH model is needed. Future research should examine how the four identified conditions could vary across industries, learner demographics and organizational cultures. L&D professionals should map learning objectives before selecting platforms, embed human touchpoints and prioritize outcome metrics like longitudinal tracking, behavioral transfer and supervisor assessments over completion rates. AI-driven learning risks reinforcing digital inequalities and historical biases in data. Organizations should ensure digital fluency, AI governance and conduct cultural diagnoses before deployment to prevent marginalization of digitally excluded employees. Grounded in SLT, the HMH model offers a novel diagnostic framework for understanding why high completion rates often fail to translate into performance improvements and for helping organizations assess where AI-mediated learning will succeed.