Access to high-quality vocal music education is not equitable. Many students in rural and under-resourced schools do not have trained music teachers and face limited school funding (State Education Agency Directors of Arts Education, 2019; Rabkin & Hedberg, 2011). This study looks at how artificial intelligence (AI) singing applications may help reduce these gaps by giving students adaptive feedback, pitch detection, and guided practice (Luckin et al., 2016; Roll & Wylie, 2016). It is guided by constructivist learning theory, which understands learning as active and scaffolded (Vygotsky, 1978; Bruner, 1966), and by equity frameworks such as culturally relevant and sustaining pedagogy (Ladson-Billings, 1995, 2006; Paris & Alim, 2014). The main research purpose is: How can AI singing applications support student engagement, skill development, and more equitable access to vocal music learning in different educational settings? This qualitative case study will follow three groups of students (ages 12-18) during one semester (10-12 weeks). Data will come from pre- and post-questionnaires, weekly journals, facilitator observations, focus groups, and teacher interviews. Students will practice with the app two to three times each week, while facilitators observe their learning and difficulties. Thematic analysis will be used in an iterative way (Charmaz, 2006; Braun & Clarke, 2006). This project adds to the small but growing literature on the use of AI tools in music education (Bauer, 2014; Crawford, 2017; Holmes et al., 2019; Selwyn, 2019). By listening to student voices and comparing public school, after-school, and online contexts, the study will show both the possibilities and the limits of AI as a tool for reducing inequities in vocal music education. The expected results will give practical ideas for more inclusive, technology-based pedagogy, and will also add to wider debates on educational justice (Warschauer & Matuchniak, 2010). By the time of the ISME 2026 conference, partial data collection and preliminary analysis will be completed, allowing for the presentation of early insights alongside the study design.