Supporting Without Replacing: Designing an AI-Driven Companion for Children with Dyslexia
Future Software Technologies
Semester programme:Master Applied IT
Research group:Interaction Design
Project group members:Thomas Dielesen
Margit Kiewit
Evelien van de Garde-Perik
Mark de Graaf
Project description
Children with dyslexia often require structured practice, repetition, and continuous feedback to develop reading and spelling skills. Due to limited educational resources and teacher shortages, providing individualized support remains challenging. This project investigates how an AI-driven companion can support children with dyslexia during reading and spelling exercises while remaining within appropriate educational and ethical boundaries.
Using a Research Through Design approach, the project combines educational research, interaction design, ethical analysis, and prototype development. The resulting AI companion provides guidance, feedback, and motivation during learning activities. The project explores how teaching practices can be translated into AI-supported interactions without replacing the role of teachers.
Context
The project is situated within the domains of educational technology, human-computer interaction, and artificial intelligence. It focuses on supporting children with dyslexia, a learning difficulty that affects reading and spelling development.
Recent advances in conversational AI and speech technologies create opportunities for new forms of educational support. At the same time, increasing teacher shortages place pressure on the availability of individualized guidance for students with learning difficulties.
The project explores how AI-driven companions can support structured educational activities such as guided practice, feedback, and repetition. Particular attention is given to motivation, interaction design, and the ethical boundaries of AI in educational contexts. Rather than replacing teachers, the AI companion is positioned as a support system that can assist with specific instructional activities while teachers remain responsible for educational decision-making and long-term guidance.
The project is conducted within a research environment focused on interaction design and AI-supported learning experiences.
Results
The project resulted in a validated design framework for an AI-driven educational companion aimed at supporting children with dyslexia during reading and spelling exercises.
The research identified which instructional tasks can be effectively supported by AI, including guided practice, repetition, immediate feedback, and structured learning activities. It also identified educational responsibilities that should remain with teachers, such as diagnosing learning difficulties, interpreting long-term progress, and making educational decisions.
The findings were translated into personas, interaction flows, feedback strategies, ethical design requirements, and multiple prototype iterations. Validation with educational specialists resulted in several design improvements, including reflective feedback moments, confidence-recovery mechanisms, and exercise structures aligned with existing educational practice.
The final outcome is a functional prototype that demonstrates how AI can support structured learning activities while remaining within clearly defined ethical boundaries. The project contributes insights into the design of AI-supported educational systems and provides a foundation for future research into motivation, engagement, and human-AI collaboration in education.
About the project group
This project is conducted individually as part of the Master Applied IT programme at Fontys ICT. The project is carried out full-time during the semester and combines research, interaction design, and software development. The work follows a Research Through Design approach, involving iterative cycles of research, prototyping, validation, and refinement.