Building on research and practice in creative music education (Burnard & Murphy, 2017), this study sought to explore the potential for generative AI to leverage creative music-making ideas among preservice students in early childhood education and to develop critical, analytical and creative skills among the students in the undertaking of the work. While the use of AI in education contexts raises valid concerns in relation to the originality of ideas, cultural bias, ethics (Vieriu & Petrea, 2025) and not least, unchartered dimensions (UNESCO, 2023), a transparent approach can place an explicit focus on such matters, recognising both the limitations and potential benefits. In this study, the adoption of a practical, problem-solving framework, and with a view to extending creative possibilities in education (Glăveanu, 2022), students are positioned on the nexus of past and future knowledge, as thoughtful, agentic learners who have the capacity to extend their own understandings of theory and creative practice relevant to music education in early years through use of AI. At the same time, they learn to do so in a supervised and safe way that is transparently undertaken. The study outlines how the students experiment with AI and a selected set of practical objects to explore their potential for musical play in early years settings. The process involves an integration of theory and practice, including procedural and ethical issues. In keeping with the frame of possibility thinking, findings reveal how the study has opened up dialogue on childhood (musical) cultures (Sheehan Campbell, 1998), debate on creative engagement with music education, ethnomusicological perspectives on musical instruments and their sound potential, as well as heightening students’ situational interest (Guo & Fryer, 2025) in both the process and its transfer to settings.