Empowering Visually Impaired Students to Explore Mathematical Tactile Graphics Using AI: MathTalk

Blind students frequently encounter significant challenges when attempting to explore mathematical tactile graphics, especially those involving special complexity, overlapping elements, and abstract concepts. Traditional tactile materials are static and non-interactive, requiring students to rely heavily on teachers or peers, which makes it harder for them to learn on their own and truly understand the concepts. MathTalk, an AI-powered voicebot designed to provide real-time, step-by-step audio instructions that scaffold students through tactile STEM graphics. Developed using the Design and Development Research (DDR) model, MathTalk have three key phases: needs analysis through qualitative interviews with blind students in Malaysia, expert validation using the Fuzzy Delphi Method involving specialist in special education and AI and experimental evaluation with 19 secondary school blind students. Results indicate that a significant majority of students demonstrated improved in understanding, quicker tactile exploration, and increased autonomy in the exploration. The system operates via a simple 8-key tactile interface by using mini keyboard, ensuring ease of use without requiring visual interfaces. Beyond its pedagogical value, MathTalk is scalable, cost effective, and aligned with national education policies that empower inclusive and technology-driven learning environments. Its commercialization potential is strong, with applications in schools and NGOs that support visual impaired communities. Mathtalk initiates transformative leaps in accessible STEM education, offering blind students and an equitable and empowering learning tool.