Multimodal generative AI for pre-practicum mental health support competency development: a human-supervised framework for education–childcare partnerships

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Generative artificial intelligence (GenAI) and multimodal interaction are attracting growing attention in mental health-related education and care. At the same time, pre-service childcare professionals are increasingly expected to observe and respond sensitively to infants’ and toddlers’ emotional signals, attachment needs, separation reactions, and interaction patterns. These contextual and ethically sensitive capacities require more than lectures or procedural training. Authentic practicum remains indispensable, but safeguarding, consent, privacy, placement availability, supervisory capacity, and the unpredictability of naturally occurring situations can limit safe rehearsal. Drawing on research on large language models, conversational agents, responsive caregiving, competency-based assessment, work-integrated learning, implementation science, and health AI governance, this conceptual analysis proposes five design domains for education–childcare partnerships: competency mapping, multimodal simulation, AI-supported reflective feedback, practicum readiness assessment, and partnership governance. Multimodal GenAI is positioned as human-supervised learning infrastructure that may extend established simulation through adaptive scenario variation, contingent interaction, multimodal cue integration, and scalable formative feedback. The proposed architecture includes a developmental-fidelity gate under which synthetic infant- or toddler-like materials would be reviewed against predefined anatomical, affective, temporal, cultural, and pedagogical criteria before use. The framework defines non-clinical mental health support competence through five observable dimensions and specifies human authority, data-minimization boundaries, and a staged research pathway beginning with a candidate Pre-Practicum Mental Health Support Competency Assessment (PP-MHSCA) blueprint. The proposed competency levels, readiness categories, critical-error rules, scoring procedures, and thresholds are candidate specifications requiring stakeholder review, pilot testing, standard setting, and empirical validation before implementation. Comparative studies with conventional simulation and lecture/case preparation, followed by prospective tests of practicum performance, can determine whether the approach adds educational value. The framework offers a cautious and governable bridge to practicum without displacing educators, mentors, or real-world experience. The framework is grounded in broadly applicable professional standards and governance principles, with China proposed as the initial context for implementation and empirical evaluation. Evidence generated through this process is expected to inform the framework’s subsequent adaptation, transfer, and application in other national and regional contexts.