NAO Robot RCT — Boosting Social Communication for Autistic Children in Hong Kong

Autism-specific evidence is thin in most AI-in-education evidence bases, and much published research on social robots for autistic children relies on small, uncontrolled case studies. Researchers at The Education University of Hong Kong (EdUHK) set out to test a social-robot intervention against a proper comparison design rather than case reports alone.

The trial aimed to establish whether a NAO humanoid robot delivering a structured social-skills curriculum could produce greater communication and social-interaction gains for autistic children than either human-led instruction or no intervention.

Sixty children with autism spectrum disorder, aged 5 to 11 (IQ of at least 70, able to follow simple commands), were randomised into three parallel arms: a robotic-intervention group taught by a NAO humanoid robot, a human-instructed group receiving the same curriculum from a person, and a no-treatment control group. Over 12 weekly sessions, children in the intervention arms worked through structured social games, story-based activities and singing-and-dancing routines covering two-way communication, basic emotions, imitation and reciprocal responses. Outcomes were measured with the Autism Diagnostic Observation Schedule (ADOS) plus parent-reported change, and the trial was pre-registered on ClinicalTrials.gov (NCT04879303).

The robotic-intervention group showed significantly greater improvements in communication and reciprocal social interaction than either the human-instructed or control groups, on both the performance-based assessment and parent report. The results were published in the peer-reviewed Journal of Autism and Developmental Disorders (2025).

The trial is single-site with a modest sample (60 children) and a short, 12-week intervention window, so the authors call for larger, longer, multi-site replication before the approach is treated as established practice. NAO hardware costs and the need for trained facilitators are also practical constraints on near-term scale-up in typical classrooms.