A Qualitative Exploration of Parent Engagement and Challenges with AI in ECCE Homeschooling

Abstract

Early Childhood Care and Education (ECCE) is at a turning point, where play-based learning meets with generative and conversational artificial intelligence (AI). The present study examines the current practices of parents in integrating AI tools (ChatGPT, Gemini, Meta AI, Google Assistant, MagicSchool.ai and Khan Academy AI) in ECCE level homeschooling in Punjab, Pakistan, and examines the practical, pedagogical and technical challenges. A qualitative exploratory design with an interpretivist paradigm was used, and semi-structured interviews were held with a sample of 25 home schooling parents in four districts (Lahore, Rawalpindi, Gujrat and Multan). A hybrid analytical approach of reflexive thematic analysis using NVivo 14 and text analytics using python was adopted. Results show the parents have systematically adopted AI as an ‘adaptive scaffolding’ tool to personalize their instruction, create personalized curriculum content and find interactive resources, which increased children's learning engagement and self-regulation. Yet, there are limits to successful integration, including the absence of stable infrastructure (electricity supply and internet connectivity), subscription issues, Urdu language limitations, and strong ethical-cultural gaps between AI-generated products and the local curricula (e.g., Pakistan's Single National Curriculum). Of fundamental importance, the study finds that the parent is the first 'Interaction Variable', as they are required to engage in 'Parental Pedagogical Mediation' to protect against screen dependency, and to maintain a concrete and balanced Montessori inspired learning environment. Finally, a synthesized Human-Centered AI-Supported Homeschooling Framework is proposed, which demonstrates the pedagogical viability of AI is not dependent on the sophistication of the technology but rather structured and empathetic human mediation.