The use and significance of artificial intelligence (AI) have now become central topics in educational discourse (De Witt et al., 2023). Teachers are seen as responsible for enabling students to engage with AI in a “critical, constructive, and reflective manner” (KMK, 2024). Teachers should not only actively work with AI themselves, but also understand its technical foundations and be able to assess its opportunities, risks, and legal framework. Its use should be “open, critical, and responsible.” This also results in new requirements for music teachers in particular. However, there has been little empirical research to date on how music teachers deal with the specific challenges of integrating AI into their teaching practice (cf. Buchborn & Treß, 2023; Cheng, 2023; Treß, i.E.). In this presentation, we focus on the perspectives of music teachers and examine the norms, everyday theories, and implicit knowledge bases that underlie their use of AI in their professional practice. The data for our study is based on group discussions with music teachers that were conducted as part of a teacher training course. These are interpreted using the documentary method based on praxeological sociology of knowledge (Bohnsack, 2018). The aim is to examine the implicit, action-guiding knowledge as well as the everyday theories and norms of the participants with regard to the role of AI. Initial results point to several key themes: Participants recognize the benefits of AI for tasks such as literature research and text summarization, but still have concerns about accuracy and reliability, especially when it comes to evolving information. While some teachers are motivated to explore AI tools and see them as a helpful instrument for promoting creativity, for example, others are more distant and cautious. They question the depth and originality of the results obtained through AI and express ethical and legal concerns. Explicit assessments of AI also vary: positive assessments of work efficiency and time savings are offset by concerns that over-reliance on AI tools could hinder critical thinking and independence or threaten the ‘creative core’ of the subject of music.  Our findings point to a tension between recognition of the relevance of AI and skepticism about its significance in music education, as well as between teaching practice and pedagogical norms, as illustrated at the outset by the example of the KMK recommendations for action.