Navigating Assistive Robotics for Autism. Assessing Effectiveness, Pros, Cons, and Future Pathways

Abstract Assistive robotics holds promise for supporting therapeutic and educational interventions for autistic children, yet key questions remain regarding its practical benefits, limitations, and ethical implications. This study critically examines a 4-month therapeutic laboratory in which a humanoid robot (Pepper) was integrated into weekly sessions with four autistic children (ages 11–14, level 1 support needs). Through a mixed-methods approach, combining behavioural observations, standardised questionnaires, and thematic analysis of video recordings, we evaluated the robot’s role in fostering autonomy, communication, and engagement. Quantitative results showed limited improvements in functional autonomy, while qualitative findings highlighted the robot’s potential to stimulate curiosity, collaborative behaviours, and social participation, especially through unexpected modalities such as internet meme references. Challenges emerged around conversational rigidity, limited adaptability, and user frustration due to breakdowns in feedback and interaction flow. The study underscores the value of participatory, co-designed robotics in therapeutic contexts and offers design-oriented recommendations for developing more responsive, inclusive, and culturally adapted robot behaviours in child-centered interventions.