Invisible in the Music Classroom: AI as a Support Tool for Students with Dyscalculia

Dyscalculia represents a paradigmatic case of pedagogical invisibility in educational contexts. Despite growing familiarity with the term among educators, research continues to document significant knowledge gaps and persistent misconceptions about this condition (Roulstone et al., 2026): affected students are often misjudged by teachers as lacking motivation or effort, rather than as expressing a specific cognitive vulnerability (Schul- te-Körne, 2016). This dynamic becomes particularly pronounced in mu- sic education settings, where dyscalculia has received limited scholarly attention, and teachers often lack practical tools to identify and support affected students. Indeed, the specific challenges of dyscalculia risk being confused with lack of engagement or emotional factors such as stress and performance anxiety, making it harder for music educators to recognize and address them effectively. Individual practice, therefore, represents a privileged context for observing these difficulties in their most authentic form, away from the pressure of the lesson setting (Dos Santos Silva et al., 2025). Against this backdrop, this study describes the integration of an AI-based intervention into individual practice and examines the changes observed in students’ learning processes and in teachers’ pedagogical aware- ness, within a complex educational context that also included other sup- port strategies. This qualitative case study focuses on a small group of string players with dyscalculia who participated in a didactic intervention using AI-based software during individual practice. Data were collected from the AI application logs generated during practice sessions and from students’ practice diaries. Video recordings of each lesson provided an additional source to track changes in performance over time. The findings suggest that changes were observed on two levels. On the one hand, AI applica- tion logs made students’ difficulties visible to teachers, allowing them to better identify students’ specific needs and respond more effectively during lessons. On the other hand, video recordings and students’ practice diaries revealed a progressive improvement in performance quality, characterized by fewer errors, greater fluency, and a growing sense of confidence during performance. This paper contributes to the interdisciplinary debate on in- visible vulnerabilities and AI-mediated educational innovation, document- ing how AI tools can make visible otherwise invisible learning processes, simultaneously supporting students with dyscalculia and enhancing teach- ers’ pedagogical awareness in a context — instrumental music education — where this fragility has long remained without adequate recognition or support.