Serious Games for Education and Intervention in Autism Spectrum Disorder: A Scoping Review of Learning Theories, AI, VR, XR, and Robotics

Serious Games (SGs) have increasingly been investigated as tools for education and intervention for individuals with Autism Spectrum Disorder (ASD), while emerging technologies such as Artificial Intelligence (AI), Virtual Reality (VR), Extended Reality (XR), and Human–Robot Interaction (HRI) are expanding their potential for personalization, immersion, and adaptive interaction. However, the literature remains fragmented across pedagogical theory, Serious Game design, and technology-focused interventions, making it difficult to understand how these dimensions are connected within the ASD context. This scoping review therefore maps the available literature on Serious Games for individuals with ASD, focusing on two complementary questions: (RQ1) Which learning theories, pedagogical approaches, and educational or developmental outcomes underpin the design and implementation of Serious Games for individuals with ASD? and (RQ2) How are intelligent and immersive technologies, including AI, VR, XR, and HRI/robotics, integrated into Serious Games, and what benefits and challenges are reported? A structured literature search was conducted across Scopus, ERIC, Web of Science, and the ACM Digital Library, with Google Scholar consulted as a supplementary source, covering publications from January 2005 to June 2025. Following duplicate removal, title/abstract screening, and full-text assessment, 40 studies constituted the formal evidence base. The evidence was organized through descriptive and thematic synthesis, with studies mapped according to theoretical foundations, Serious Game or intervention characteristics, technological modalities, and reported outcomes. The review identified substantial theoretical diversity, including constructivism, socio-cultural theory, social learning theory, self-determination theory, behaviorism, developmental approaches, and intervention-oriented frameworks. AI was mainly associated with personalization, learner modeling, behavioral analysis, and adaptive feedback, whereas VR/XR primarily supported immersive and controlled environments for experiential learning and skills practice. Robotics and HRI were associated with structured and embodied interaction. Reported challenges included accessibility, cost, privacy, cybersickness, methodological heterogeneity, and limited evidence concerning transfer and maintenance of skills. Overall, the review provides a structured overview of the field and indicates priorities for further empirical investigation.