Brief Report: Development of a Robotic Intervention Platform for Young Children with ASD.

Warren et al. (2015) built a small robot that watches a child and gives live feedback. The robot copies simple arm and hand moves. If the copy is good, the robot smiles and lights up. If not, it tries again.

Ten preschoolers with autism and ten typical peers each played two games. One game used the robot. The other used a human adult doing the same moves. The team filmed which version kept kids looking and copying best.

Kids with autism looked longer and copied moves better when the robot led the game. Typical kids did well with both partners, but the robot still held their gaze.

The robot's live feedback seemed to help the autistic group stay in the game.

Kumazaki et al. (2018) ran a similar lab test and saw the same thing: simple robots and even on-screen avatars pulled more looks from 3-year-olds with autism than a real person. This backs up the attention boost.

Yet Mruzek et al. (2019) adds a warning. They also got more face gaze with a robot, but joint attention to a shared toy did not improve. More looking does not always mean deeper shared focus.

Annunziata et al. (2024) took the idea further. They gave 10 Italian preschoolers 14 weeks of robot mirroring games. Only 4 kids moved past the warm-up, but those 4 showed social gains six months later. The early robot spark can grow if you keep the program going.

If you run early-intervention sessions, try starting with a short robot or animated avatar demo to grab attention. Once the child is looking, switch quickly to shared play so the gaze turns into real joint attention. Track which kids stay engaged and plan longer robot-led mirroring blocks for them. The tool is not magic, but it can open the door.