As a musician and a teacher, I am becoming increasingly interested in the role of Artificial Intelligence (AI) in the primary music classroom in all its forms. While AI has primarily been associated with subjects like science, mathematics and languages, its potential to be used in music education is becoming significant. In primary schools, where creativity, collaboration and engagement are central, AI tools can offer exciting opportunities and challenges for students and educators alike.
In this article, I hope to outline an argument for how AI can both support and hinder the music classroom so that you can consider how it can best work for you – because, let's face it, it's not going anywhere!
Mixed-ability teaching
One of AI's key strengths is its ability to personalise learning. Primary music classrooms often include students of varied musical abilities, from complete beginners to those with prior musical experience. AI can provide tailored learning experiences for each student, adapting to their progress and needs. For instance, platforms like Yousician use AI to monitor progress, provide instant feedback on performance and adapt lessons to adjust the difficulty level of lessons. It has elements of ‘gamification’, which encourages flow learning and increased motivation. The app can support teachers in tailoring lessons according to where a student is and will provide in-the-moment feedback to support the student's progression, becoming the ultimate adaptive teacher. If you are teaching whole-class ukulele, guitar or keyboard, this could be a way to monitor progress and allow students to practise at their own pace.
However, while AI can enhance learning, it cannot replace the value of human interaction and mentorship. If a group of children needs dedicated ‘human’ input to access the instrument successfully, you could spend time with them, knowing that AI is supporting other learners and helping them progress while you do this.
Simplifying composition
AI tools can help foster students' creativity by simplifying the process of music composition. Apps like Soundation and GarageBand allow young learners to experiment with creating their own music. Soundations' AI suggestions help students experiment with different genres and sounds, offering them new ideas as they work. The app also provides instant feedback on rhythm and pitch. Music hubs such as Derbyshire's and various schools around the country signpost children towards this platform to experiment and create in their own time. Soundation gives students with minimal musical backgrounds the opportunity to create professional-sounding tracks, helping them build confidence and creativity and feel connected to classroom music in a new way. However, using these tools, students might lose valuable conventional music skills if relying too heavily on AI. For example, they might not develop the fundamental concepts of music notation, understand structure and form, grasp harmonic understanding, or develop their sight-reading skills. AI platforms allow students to bypass these areas for more automated, intuitive music-making.
Teacher short cuts
Apart from AI suggestions and tools already embedded into apps, AI assistants such as ChatGPT, Microsoft Copilot, and Meta AI hold the potential to be helpful for the lone music teacher or music lead in a primary school. I asked two of these AI assistants to produce a song with a four-week lesson plan for Year 4 and a xylophone part to accompany it. I asked the following questions with my best manners because when AI takes over, I want it to think of me favourably!
- Please create a song about friendship for children in Year 4 in the UK.
- Please create a simple xylophone part (to be played by the children) to accompany the song.
- Please use stave notation for the xylophone part.
- Please create a four-week lesson plan based on this song for a Year 3 class with learning objectives taken from the Model Music Curriculum.
- Please generate a piano part for a Grade 3 pupil in my class.
What came back in 0.45 seconds was better than I could have created in that time. The song was basic, and had no information about the melody, tempo, style and so on. But it had a solid rhyming structure, and the content (shown in Fig.1) was appropriate for young children. The notation was presented in code for MusicXML (Fig.2), and through a bit of Googling and with the help of my tech industry husband, I was able to put this into Sibelius and Musescore (Fig.3). This wasn't as successful as I had hoped, but AI is still learning and the basic concepts are there. I could have refined the prompts, and maybe this would have improved the outcome; but I could probably write out a xylophone part in less time than finding the right prompts for AI to generate it. The lesson plans were basic, but they provided a structure or a decent starting point for me to develop my ideas (Fig.4). With further prompt engineering, I'm sure you could effectively generate plans and music quickly in collaboration with AI tools. AI is continuing to evolve; it is learning through every prompt and we can be sure that what it can't do now, it will be able to do tomorrow.
Fig.2 MusicXML format for creating stave notation
Fig.3 Converted MusicXML
Fig.4 Proposed lesson plan
It is interesting to go through this process with your students. Using AI song-generation sites such as SUNO can show children just how quickly a song can be generated. They can create in the style of their favourite artist, all with a few lines of prompts. The more specific they make the prompts, the increased complexity and detail in what is produced. Maybe this will become a new route for the creative process.
Summary
AI holds tremendous potential to enrich music education in primary schools, reduce your workload and create a balanced, engaging and inclusive learning environment when used as alongside traditional methods, as a supportive tool. Ultimately, it cannot, I feel, replace the intrinsic human qualities of music-making, such as playing together, improvisation, expression, and collaboration.
