Research Paper Proposes Relational AI Framework for Education Based on Indigenous Worldviews

A new research paper argues that current generative AI applications in education risk undermining relational learning processes by prioritizing efficiency and individualization. Published on arXiv as preprint 2604.19099, the work critiques how AI in education (AIED) research has failed to articulate designs that sustain the social and ecological relationships essential for learning. The authors propose reframing learner-AI interactions as context-specific relationships with clear purposes and boundaries, rather than as replacements for human interaction. This relational approach is grounded in participatory design practices and draws inspiration from Indigenous worldviews, including Aboriginal Australian perspectives. The paper emphasizes that education is fundamentally a social, constructive, and relational practice, not merely information transmission or performance optimization. It addresses the growing adoption of generative AI while calling for designs that support rather than weaken relational learning.