Abstract for: Improving Accessibility of The System Dynamics Learning Guide Using the 'Systems Thinking for the Visually Impaired' Custom GPT
This paper presents a case study on developing a chatbot to improve accessibility of the System Dynamics Learning Guide (Deaton & MacDonald, 2025) used at James Madison University to teach an introduction course in Systems Thinking. Individuals with visual impairments face significant challenges in understanding Systems Thinking representations due to their visual nature. This research explores the use of Generative AI (GenAI) to enhance comprehension while making educational materials more inclusive. The Systems Thinking for the Visually Impaired chatbot (Systems Thinking for the Visually Impaired, 2024) uses AI-driven image recognition and structured prompts to generate accurate, text-based descriptions of Systems Thinking representations. Designed for users with visual impairments, this chatbot aims to enhance comprehension by providing clear and contextual descriptions of these models. Initial results using images from the Learning Guide (Deaton & MacDonald, 2025), indicate that the chatbot effectively generates accurate and spatially aware descriptions. The structured prompt framework helps maintain consistency and clarity in responses. By translating visual models into detailed text descriptions, this chatbot demonstrates the potential of AI-driven tools in improving access to educational materials for students with visual impairments. This study highlights the potential of AI-driven tools in advancing Systems Thinking in STEM education. By improving access to traditionally visual Systems Thinking representations, the chatbot aims to support visually impaired learners and offers educators a scalable solution to enhance teaching materials. The results demonstrate that integrating AI-driven tools can promote inclusivity and broaden educational opportunities. Used for the development of the image recognition prompts and for improving clarity