Examination of Artificial Intelligence Language Models Developed for Piano in Terms of Their Technical and Educational Aspects

Abstract

With the widespread adoption of artificial intelligence tools in our daily lives in recent years, we have seen the development of sub-language models specialized in various fields. It is also evident that these specialized language models have been developed for music and music education. In this regard, this study aims to examine piano-focused AI models found in the current literature under eight main categories—developer/platform, year, primary purpose, target audience, pedagogical approach, interaction status, output quality, and accessibility—and to generate insights regarding their applicability in education. The research is a descriptive analysis study designed using the document review model from qualitative research methods. It is possible to create different specialized language models tailored to specific purposes using the infrastructure of large language models. These language models can be created either through professional coding or via a prompt-based approach. Domain experts can also develop sub-language models based on prompts according to the professional requirements of their field without the need for a software developer’s support. In this regard, OpenAI’s ChatGPT model, with its prompt-based structure, possesses a large number of sub-language models tailored for piano education. For this reason, the scope of this research is limited to the use of ChatGPT’s sub-language models for piano education. These sub-language models will be analyzed using evaluation criteria developed with input from experts. These criteria include items addressing both the technical characteristics of these language models and their applicability in education. The findings obtained from the analysis will be evaluated in light of these criteria. Recommendations will be provided regarding the requirements of piano education and how these models can be utilized in piano education.