Intergrating AI with Social Robots to support children with ASD
The Social Robotics Lab at Jomo Kenyatta University of Agriculture and Technology (JKUAT) is once again proving to be a hub for impactful innovation. Margaret Gichuhi, a Master’s student in Software Engineering, is currently developing a prototype AI-driven speech therapy solution tailored for children with Autism spectrum disorder (ASD). Her research delves into how social robots, when integrated with generative AI, can transform therapeutic approaches by delivering structured and responsive sessions using Pepper, a humanoid robot.
Children with ASD often face delayed or impaired speech development, requiring long-term specialized therapy. Unfortunately, traditional approaches are often expensive, hard to access, and lack personalization. Margaret’s system, still in its trial phase, addresses these gaps through a robot-assisted platform powered by GPT, DALL·E, and Whisper.
Supervised by Dr Eunice Njeri and Prof. Stephen Kimani, Margaret’s research offers a fresh approach to therapy. By programming Pepper to interact in predictable, non-judgmental ways, the system provides comfort and structure, qualities that children with ASD respond well to.
Although the system is still in its developmental phase and has not yet been tested in real-world environments, it is functional in simulated settings. The solution incorporates several advanced elements: Pepper serves as the interactive interface for both children and therapists, using its touchscreen to display visuals and activities, while its speech functions facilitate guided verbal exercises.
Why Do Children with Autism Respond Well to Pepper, a Social Robot? Pepper has been specially programmed to interact in predictable and non-judgmental ways, something that children with ASD often find comforting.
The robot provides structured, consistent interactions, which are effective in both therapeutic and educational settings. By integrating Generative AI, Pepper becomes more than a scripted robot. It can adapt language complexity, maintain conversational flow, and offer feedback, creating a safe space for practicing.
This project aligns with JKUAT’s commitment to innovation that serves real-world needs, showing how robotics and AI can promote connection, learning, and inclusion. Margaret’s work highlights the power of local talent to build tools that are contextually relevant, ethically sound, and inclusive.
