Roles, Strengths and Challenges of Using Robots in Interventions for Children with Autism Spectrum Disorder (ASD).

The team ran focus groups with therapists, teachers, and parents. They asked how the small humanoid robot KASPAR was used during autism sessions.

From the talk they sorted every use into six job titles the robot could hold. They also listed eight pluses and four headaches that came with the metal coworker.

KASPAR can wear six hats: provoker, reinforcer, trainer, mediator, prompter, and even diagnostician. Each hat matches a clear clinical move you already do by hand.

The big wins were steady attention, safe practice, and fun. The big pains were slow response time and weak carry-over to new places or people.

Butler et al. (2021) and Callahan et al. (2023) cut stereotypy in half with plain DRO or RIRD schedules—no wires needed. Their fast, large effects seem to clash with KASPAR’s mild, spotty gains. The gap is method: the robot study is qualitative, the DRO/RIRD studies are single-case experiments targeting one behavior.

Cicchetti et al. (2014) warned that computer training for autism often stays on the screen. KASPAR’s team spotted the same leak: kids talk to the robot, then freeze with humans. Both papers shout, “Plan for real-world transfer.”

Silva et al. (2019) offer a fix: build Persona families before you buy gear. Map the child’s likes, fears, and daily stages, then pick the robot’s role. The six-hat list gives you ready-made roles to plug into those personas.

You can borrow KASPAR’s six job titles even without the robot. Use a puppet, tablet, or peer to be the provoker or mediator. Whatever toy you pick, add generalization probes from day one: swap rooms, adults, and materials so the skill travels past the shiny object. And if you trial a robot, budget extra sessions for human-to-human transfer; the cute face alone won’t do it.