Large language models (LLMs) are reshaping how music educators search, write, plan, and reflect. In music education, they are already used for lesson planning, resource curation, assessment design, strategy, research and many other applications (Cheng, 2025; Merchán Sánchez‑Jara et al., 2024). LLMs rapidly generate options at scale, reformulate text and other media and adapt to constraints in seconds. However, LLMs respond to training data, available resources and the precision, structure and content of language input.This paper/presentation explores the conception of large language models as conditional learning spaces that provide a means for co-constructed, visible and exploratory dialogue for music teachers. I do this to lay the groundwork for a policy analysis. First, I explore the structural and logical limitations of prompt behavior as revealed through research on prompt engineering (Bsharat et al., 2024; Bubeck et al., 2023; OpenAI, 2023, 2025; Sahoo et al., 2025), identifying strategies that enhance reasoning quality and consistency. Prompt behavior is considered from multiple perspectives, including producing desired outputs, questioning assumptions, providing affordances and limiting cognitive capacity, which is then contextualized to intellectual work in music education, i.e., curricular planning, materials developing, administrative work, brainstorming tasks, conceptual planning, etc.). Third, I root a policy problem (Barduch, 2024) in the contradiction that LLMs appear to have considerable potential for accelerating and enhancing music education work (Cheng, 2025); however, this may require overriding natural inclinations and default templates currently being built into technology applications. Additionally, I discuss impacts for human well-being and the environment (US GAO, 2025). Fourth, I develop two alternative visions, one in which LLMs are embedded into music teacher education to enhance teaching and learning and another in which LLM use in music teacher education is discouraged. Fifth, I draw upon current LLM institutional guidelines (USDOE, 2023; NAfME, 2025; UNESCO, 2025a, 2025b) to examine the discourses that may be meaningful for shaping thinking about these effects for music teacher education. As a whole, the paper seeks to deepen more intentional, equitable, forward-thinking and critical dialogue on large language models as part of music education professional practice.