Nineteenth International Conference on Technology, Knowledge, and Society

Abstract

While the notion of intelligence within human cognition is expansive, artificial intelligence (AI) often invokes imagery and computer models of neurotypical human thinking. The implicit cognitive assumptions of AI reinforce power dynamics around the human experience of acquiring knowledge. This paper considers the use of machine learning (ML) to increase legibility of neurodiverse thought through a pilot study on number-space synesthesia. Using Google Teachable Machine (GTM) software, the study develops an algorithm to identify numeric expression from a perspective of number-space synesthesia. The model can be used as a tool to bring neurodiverse methods in mathematics for AI engineering, envisioning what inclusive AI design can look like.