Oxytocin is a potential therapeutic for the core symptoms of autism spectrum disorder (ASD), which is currently untreatable with pharmacotherapy. Previous clinical trials of a single dose of oxytocin have consistently reported significantly positive effects on various experimental measures associated with the core symptoms of ASD. These studies used various experimental measures as surrogate endpoints of the trials. However, to date, randomized clinical trials of continual administration of oxytocin have failed to reveal significant positive effects on clinically meaningful endpoints, such as how those with ASD interact during interpersonal interactions. This article reviews both the negative and positive effects of oxytocin on the core symptoms of ASD and their surrogate markers. Some unresolved and critical issues on the development of oxytocin as a new therapeutic have been extracted: optimization of dose, duration of oxytocin treatment, and the development of objective and reliable measurements of clinically meaningful endpoints for the core symptoms of ASD. Furthermore, optimization to the intranasal delivery system and careful consideration of how individuals respond differently to treatments should be addressed in future studies.