Prompt-Mediated Curriculum Development: a human–AI co-design framework for AI-mediated curriculum development in higher education

The integration of generative and agentic artificial intelligence (AI) into higher education is transforming curriculum development by enabling AI systems to contribute to the design of learning outcomes, assessments, instructional activities, and other curriculum artefacts. Despite growing interest in AI literacy, prompt engineering, and human–AI collaboration, limited attention has been given to how educators enact curriculum expertise within AI-mediated design environments. This conceptual paper introduces the Prompt-Mediated Curriculum Development (PMCD) framework, which conceptualises curriculum development as an iterative process of human–AI co-design mediated through prompt literacy. Drawing on curriculum theory, sociocultural perspectives on mediation, human–AI collaboration, and AI literacy scholarship, the framework positions prompt literacy as the central mechanism linking educator expertise, agentic AI systems, and curriculum artefacts. Prompt literacy is defined as the ability to formulate, refine, and evaluate prompts in ways that support pedagogically aligned and contextually relevant educational outcomes. Building on the framework, the paper proposes a Prompt Literacy Taxonomy comprising three developmental levels: Foundational Prompting, Pedagogical Prompting, and Agentic Orchestration. The taxonomy provides a conceptual basis for understanding how educators engage in sophisticated forms of AI-supported curriculum development. The paper contributes to the literature by advancing prompt literacy as a distinct form of educational expertise, proposing a theoretical model of AI-mediated curriculum development, and offering a developmental taxonomy to inform future research, faculty development, and institutional AI integration strategies.