<![CDATA[This review aims to synthesise current knowledge on AI integration in higher education and conceptualise human-AI collaboration. Using a structured narrative review of literature, it examines evolving educator roles and approaches to AI adoption. While AI can enhance productivity, concerns regarding academic integrity, over-reliance, and authentic learning remain. Fully AI-free instruction remains theoretically possible but is increasingly rare, whereas complete replacement of human educators through fully autonomous AI-led instruction is unlikely in contemporary higher education. To conceptualise varying degrees of AI involvement, a six-level taxonomy of human–AI collaboration is proposed, ranging from fully human-controlled instruction to higher levels of AI autonomy. Currently, most collaboration occurs at Levels 2 to 3, with AI serving as an instructional, evaluative, or pedagogical decision partner. Effective adoption requires AI literacy among students and educators, AI supporting rather than replacing student cognitive effort, educators designing AI-resilient tasks and assessments that prioritise reasoning and include human-in-the-loop evaluation, and institutions ensuring pedagogical alignment, ethical governance, and sustainable implementation. The taxonomy provides a shared language to characterise current practices, anticipate shifts in human–AI roles, and guide educational practice, policy, and future research.]]>
Copyrights © 2026
