Skip to content

Site Navigation

Opportunities and Challenges for AI as an Assistant in STEM Accessibility

Presenter

Markku Häkkinen PhD

Educational Testing Service, USA

Abstract

AI technologies offer many possibilities in advancing the field of accessibility, but also pose challenges and risks. One area where AI can have near term or immediate value is in STEM accessibility, and one specific area is making scientific data more accessible and inclusive. By transforming complex data into multi-modal formats, AI enables individuals with disabilities to engage more effectively with scientific content. This includes converting visual data into audio descriptions, simplifying text through natural language processing, and offering interactive, adaptive visualizations that cater to diverse needs. In choosing to apply AI, we have to recognize its limitations and seek applications and strategies for use that are effective and accurate. Treating AI as an assistant, or apprentice is crucial as it should not be seen as a perfect or ideal expert guide.  It will learn in the process of working with the student or professional.  While the potential is significant, it is crucial to acknowledge that AI technology still requires significant training and refinement. Ongoing research and development is essential to ensure these tools are effective and reliable in promoting true accessibility in STEM fields.