Artificial intelligence and epistemic interoperability: towards a sympoietic approach

Abstract

Artificial intelligence (AI) large language models (LLMs) offer transformative potential in education through their multilingual capabilities and broad accessibility to knowledge, enabling seamless access, translation, and application of information across linguistic, socio-cultural, and disciplinary boundaries. This study examines the collaborative use of LLMs by university students in Hong Kong and Australia to translate and adapt medical materials for Australian ethnic minority communities. Grounded in Donna Haraway’s concepts of sympoiesis and care, and supported by surveys, focus groups, and AI-enabled photovoice reflections, the research highlights both the opportunities and limitations of LLMs. While LLMs demonstrate powerful generative abilities, their lack of situated knowledge often requires significant human intervention to produce accurate and culturally appropriate translations. The findings advocate human-AI co-composition in fostering inclusive practices and reimagining these partnerships as sympoietic processes, emphasising mutual care and contextual sensitivity in knowledge production practices beyond tired metaphors and dichotomies of human versus machine.