Generative AI in Higher Education Instructional Design: A Systematic Review of AI Roles, Educator Competences, and Professional Development Approaches
TL;DR
An account of AI-supported ID in which the educator’s design expertise, not the technology’s generative fluency, remains the organising centre is grounds, with embedded ethics checkpoints and performance-based assessment as the best-supported design features.
Abstract
Higher education (HE) educators have rapidly adopted generative artificial intelligence (GenAI) for instructional design (ID), yet evidence on this educator-facing use remains scattered. This PRISMA-guided qualitative systematic review synthesises 78 studies (2022–2026) to examine how AI-supported ID activity is distributed across ADDIE phases, functional AI roles, and practitioner versus pre-service contexts; the competences and human-oversight practices it requires of educators; and the professional development (PD) approaches documented for building them. Documented activity concentrates in the Design phase (95% of studies), while Implementation and Evaluation remain largely unexamined. Most studies (76%) apply no named competence framework; seven competence clusters emerge, with ethical awareness and bias detection recurrent yet weakly operationalised, suggesting a possible perceptual rather than declarative deficit. Human oversight is an explicit commitment in 58% of studies but examined empirically in only 21 (27%); the review consolidates five core oversight mechanisms plus two emerging ones, for which quantified evidence remains sparse and uneven. The corpus identifies six PD approach families, with embedded ethics checkpoints and performance-based assessment as the best-supported design features. The synthesis grounds an account of AI-supported ID in which the educator’s design expertise, not the technology’s generative fluency, remains the organising centre.