Outcomes of a Robot-Assisted Social-Emotional Understanding Intervention for Young Children with Autism Spectrum Disorders.

Marino et al. (2020) ran a randomized trial with 4- to 8-year-old children with autism. Half the kids got standard group CBT. The other half got the same lessons while a friendly robot moved, spoke, and modeled feelings.

Both groups met for the same number of sessions. Therapists followed REBT steps: identify feeling, challenge thought, practice new action. The robot acted as co-teacher, not toy.

Children who worked with the robot improved more on emotion recognition, emotion comprehension, and perspective-taking. Standard CBT helped, but the robot group moved ahead.

Gains showed up on pictures of faces, story questions, and role-play tests.

EGranieri et al. (2020) pooled 18 trials and found tech-based social skills training works as well as face-to-face. Flavia’s robot trial is one of the studies inside that meta-analysis, so the new result supports the larger pattern.

Chiviacowsky et al. (2013) first showed that a robot dinosaur made kids talk more to adults. Flavia extends that idea: the robot can also teach what the words mean.

Hopkins et al. (2011) used computer avatars and got similar emotion gains. Robot or avatar, the key is the child practices with a predictable social partner.

You already run social skills groups. Adding a small robot as co-teacher may give you extra mileage without extra session time. The robot models faces, repeats phrases, and never looks bored. Kids attend longer and generalize to real faces. If your clinic owns or can borrow a humanoid robot, try plugging it into your existing CBT plan. Start with one group, track emotion scores, and compare.