Development of an AI-based composition app

On October 1, 2023, the starting signal was given for a new project in which artificial intelligence (AI) is to be combined with human intuition. The project is being led by Professor Jörn Arnecke from the Center for Music Theory at the University of Music FRANZ LISZT in Weimar. Andrew McLeod, researcher at Fraunhofer Institute for Digital Media Technology IDMT, is playing a supporting role as an advisor for the artificial musical intelligence in the project. The aim of the "Musik-Automat" project is to create new ways to learn musical composition through the interactive interplay between students and AI. An app is to be developed for this purpose.

Woman hands playing music notes
© ipopba -

Students would use the music automaton app to develop their own compositions in dialog with the app. Similar to a ChatBot, for example, the student could input some starting point for the composition. The AI would then suggest a number of possible continuations for the student to choose between, and the process would continue. Of course, the results can be listened to immediately and the AI can also make further compositional recommendations in a range of musical styles, depending on what data it was trained on.

"This AI-based composition app could be a revolution in music composition teaching," says Andrew McLeod, an expert in digital audio signal processing at Fraunhofer IDMT. The aim of music theory is to better understand composition styles and practices. By giving the students a chance to participate in the process in an interactive way, they will be forced to think comparatively about the choices, which can lead to breakthroughs in their understanding.

In addition to Andrew McLeod, the University of Music is also working on the music automation app with Australian AI expert and composer Alex James Vaughan. The combination of classical music theory approaches with modern analysis methods opens new creative possibilities. "Perhaps the music automaton app will reach people who previously had little interest in classical music and broaden their experiences in a playful way," says Ilmenau-based AI scientist and amateur musician Andrew McLeod.

The 15-month research project is funded by the Stifterverband and the Thuringian Ministry of Economic Affairs, Science and Digital Society as part of the Fellowship for Innovations in Digital Higher Education at Thuringian Universities.


Last modified: