Generative AI for STEM Learning Among Visually, Hearing, and Speech-Impaired Students

This chapter examines how generative AI (GenAI) can enhance STEM accessibility for learners with visual, hearing, and speech impairments by incorporating an integrated, multimodal design. It identifies key barriers, visual, auditory, and pace-related. It presents GenAI-driven solutions, including tactile diagram generation, data sonification, adaptive audio descriptions, math-aware captioning, sign-language avatars, and multimodal touch-sound interfaces. Design exemplars include Braille-integrated graphics, BERT-guided sonification, OCR-based chemistry reasoning, and LLM-supported molecular assembly. The chapter highlights system performance, reliability, and classroom scalability. It argues that GenAI’s strength lies in dynamically generating context-aware representations that foster synchronous participation and learner identity. Governance principles address accessibility standards, bias mitigation, privacy, and educator co-design. Ultimately, it reframes accessibility as a driver of inclusive, high-quality STEM learning for all students.