ChatGPT-Delivered Physical Activity Intervention for Children With Autism Spectrum Disorder: Pre-Post Feasibility Study.

Children with autism spectrum disorder often face barriers to regular physical activity, and families sometimes struggle to find structured, engaging ways to encourage movement at home. This small pilot trial explored whether ChatGPT, the AI chatbot, could generate physical activity suggestions that parents then delivered to their children, comparing this approach against a control group that did not receive the ChatGPT-guided activities. Parents in the intervention group found the chatbot-suggested activities feasible to carry out and reported them as interesting and useful, and children in this group showed a measurable increase in reported physical activity levels after the intervention, while the abstract does not report a formal statistical comparison between the two groups. Because this was a small feasibility study with 26 families, the findings are preliminary and mainly tell us whether this kind of digital tool can be used at all, not whether it reliably improves activity levels compared to other approaches.

If you're a parent of a child with autism spectrum disorder, you may already know how hard it can be to find activities that keep your child engaged and moving. This study looked at a new idea: having the AI chatbot ChatGPT suggest physical activities for parents to try with their children, and comparing families who used this approach to families who did not. Parents who tried the ChatGPT-suggested activities said they were doable and that their children seemed to enjoy them.

Reported physical activity levels in these children also went up after the study period. However, this was a small study of just 26 families, designed mainly to check whether this kind of tool is practical and acceptable to use, not to prove it works better than other methods.

The researchers did not report whether the increase in the chatbot group was clearly different from what happened in the comparison group. This is early, encouraging groundwork rather than solid proof, and more research with larger numbers of families would be needed before drawing firm conclusions.

This is a summary of research, not medical advice. Talk to your own healthcare provider about anything affecting your care.

This pilot randomised controlled trial addresses a practical question many physiotherapists working with paediatric autism spectrum disorder caseloads will recognise: can a widely available, low-cost digital tool like ChatGPT help parents implement home-based physical activity recommendations for their children? The study randomised 26 parent-child dyads to an application group (n=13), where parents used ChatGPT-generated activity suggestions, versus a control group (n=13), and assessed feasibility alongside physical activity levels using the Leisure Time Exercise Questionnaire. The main contribution is feasibility data: parents found the ChatGPT-suggested content usable and engaging, which is relevant information for clinicians considering low-resource digital supports as an adjunct to supervised therapy. The intervention group showed a statistically significant within-group increase in physical activity after the intervention, but the abstract does not report a between-group statistical comparison, so it cannot be concluded that ChatGPT-delivered activity outperformed the control condition. Given the very small sample size, this should be read as proof-of-concept feasibility evidence rather than an effectiveness trial, and the abstract gives no information on adverse events, long-term adherence, or whether gains were sustained.

Clinicians should treat this as an early signal warranting further controlled investigation, not as evidence supporting a specific protocol change.