Abstract
This paper reviews the application of Artificial Intelligence (AI) and Machine Learning (ML) techniques in studying autism spectrum disorders (ASD). A bibliometric analysis of 577 peer-reviewed English language articles published in 325 journals, sourced from two different databases: Scopus and Web of Science, is done with the objectives of having an overview of research, identifying relevant or impactful sources and their interconnections, assessing the contribution of AI/ML to ASD, in identifying the research gaps and mapping the future direction of research. The study finds that the USA, China, and Canada have the maximum contributions. Apparent collaborations between countries are also noticeable. The findings show that the use of AI/ML is possible in the diagnosis of ASD as evident from the many studies that have been conducted so far. Multiple ways and techniques have been explored and reported to be positive. The process is far quicker than the contemporary clinical diagnosis and offers a promising future. Perks of using AI/ML for the diagnosis of ASD include not only the possibility of low cost, easily accessible diagnosis or classification but also lightening the load on mental health professionals.