ARTIFICIAL INTELLIGENCE FOR AUTISM SPECTRUM DISORDER

Povzetek

Scoping Reviews (SR) are used in this study to investigate how AI can be used to diagnose and treat autism spectrum disorder (ASD). By examining 55 papers released between 2020 and 2024, the study sheds light on the field's developments, difficulties, and possible prospects. AI-based models, including Convolutional Neural Networks (CNNs), ResNet-18, and hybrid approaches such as GoogleNet combined with Support Vector Machines (SVM), demonstrate high diagnostic accuracy, effectively addressing limitations of traditional methods. Additionally, multimodal frameworks integrating visual, auditory, and textual data show significant promise in providing comprehensive insights into the complexities of ASD. The study also emphasizes the impact of AI in personalized interventions through technologies like Virtual Reality (VR) platforms, telehealth systems, and IoT-based monitoring devices, which improve social skills, communication, and engagement among individuals with ASD. However, notable challenges persist, such as limited access to AI technologies in resource-constrained regions, ethical concerns related to data privacy and algorithm transparency, and a lack of research focusing on adult populations with ASD. Furthermore, the absence of longitudinal studies limits the understanding of the sustained impacts of AI-driven interventions over time. This review underscores AI's transformative potential in ASD care while advocating for inclusive, ethical, and accessible solutions. Future research should focus on expanding multimodal approaches, addressing underserved populations, and validating AI applications in diverse real-world settings to enhance their impact and reach.