A TOPIC MODELLING-DRIVEN SYSTEMATIC MAPPING OF GENERATIVE AI AND CREATIVITY RESEARCH: TOWARD A CONCEPTUAL FRAMEWORK

Generative artificial intelligence is rapidly reshaping creative work, while research remains fragmented across education, creative industries, and organisational innovation. This study conducts a topic-modelling-driven systematic mapping review of 61 peer-reviewed English-language publications (2023–June 2025). Latent Dirichlet Allocation applied to titles, abstracts, and keywords identifies 20 topics consolidated into five thematic clusters: (1) creative education and pedagogy, (2) human–AI collaboration and authorship, (3) tools, platforms, and intelligent creative systems, (4) industry and cultural applications, and (5) evolving techniques and experimental insights. Cross-cluster synthesis highlights recurring GenAI roles (ideation, co-creation, personalisation, guidance/feedback), key mediators (task constraints, user capabilities, governance/design, and policy conditions), and contingent outcomes spanning fluency, originality, diversity/voice, and agency/integrity. A streamlined GenAI–Creativity Interaction Framework connects domains, roles, mediators, and outcomes to support theory-based empirical testing and responsible system design.